MSP Investigation Report 2015 - CMSC - Abitibi

Transcription

Ontario Energy Board
Commission de l’énergie de l’Ontario
Market Surveillance Panel
Report on an Investigation
into Possible Gaming
Behaviour Related to
Congestion Management
Settlement Credit Payments
by Abitibi-Consolidated
Company of Canada and
Bowater Canadian
Forest Products Inc.
Investigation No. 2010-2
February 2015
PUBLIC VERSION
2
Page Left Blank Intentionally
PUBLIC VERSION
3
TABLE OF CONTENTS
1.
EXECUTIVE SUMMARY ................................................................................................................................9
1.1
CMSC Payments ........................................................................................................................................ 10
1.2
Bowater’s Conduct ..................................................................................................................................... 11
1.3
Abitibi’s Conduct ....................................................................................................................................... 12
1.4
Observations regarding Remedial Action and Review of Continuing CMSC Payments ........................... 13
1.4.1
Remedial Action ................................................................................................................................ 13
1.4.2
Review of Continuing CMSC Payments ........................................................................................... 14
1.5
Postscript .................................................................................................................................................... 15
2.
INTRODUCTION ............................................................................................................................................ 17
3.
THE MARKET PARTICIPANTS AND DISPATCHABLE FACILITIES ................................................ 18
3.1
Bowater Canadian Forest Products Inc. ..................................................................................................... 18
3.2
Abitibi-Consolidated Company of Canada ................................................................................................ 19
4.
CMSC PAYMENTS TO BOWATER AND ABITIBI .................................................................................. 21
5.
INVESTIGATION PROCESS AND FRAMEWORK .................................................................................. 23
5.1
Market Surveillance Panel Mandate........................................................................................................... 23
5.2
Background to Investigation ...................................................................................................................... 24
5.2.1
Initial Inquiries by the MAU ............................................................................................................. 24
5.2.2
MSP Winter 2010 Monitoring Report ............................................................................................... 26
5.3
Request for an Investigation ....................................................................................................................... 26
5.4
Information Gathering ................................................................................................................................ 26
5.5
Framework for Gaming Investigations....................................................................................................... 27
6.
RELEVANT ASPECTS OF THE WHOLESALE MARKET DESIGN ..................................................... 29
6.1
Dispatchable Loads .................................................................................................................................... 29
6.2
The Two-Schedule Market Design ............................................................................................................ 31
6.3
Congestion Management Settlement Credits ............................................................................................. 34
6.3.1
The Origin of CMSC Payments......................................................................................................... 34
6.3.2
Determination of CMSC Payments ................................................................................................... 35
6.3.3
Self-Induced CMSC Payments .......................................................................................................... 38
6.3.4
Clawback of Certain CMSC Payments ............................................................................................. 39
7.
BOWATER’S CONDUCT IN RESPECT OF THE THUNDER BAY FACILITY ................................... 42
7.1
CMSC Payments to Bowater...................................................................................................................... 42
7.2
Typical Operating Pattern .......................................................................................................................... 43
7.2.1
Bidding Strategy ................................................................................................................................ 43
7.2.2
Ramping Pattern ................................................................................................................................ 44
7.2.3
Constrained-off CMSC Payments During Ramping ......................................................................... 45
7.3
Defects in Market Rules or Procedures ...................................................................................................... 50
7.4
Exploitation of Constrained-Off CMSC..................................................................................................... 52
7.4.1
Development of the Ramping CMSC Strategy ................................................................................. 53
7.4.2
Expanding the Magnitude of CMSC Using a High Bid Price ........................................................... 62
7.4.3
Submitting Maximum Bid Quantities in Excess of Consumption ..................................................... 82
7.4.4
Expanding Schedule Quantity Differences Through Ramp Down Timing ....................................... 85
7.4.5
Ramping Faster than Submitted Ramp Rates .................................................................................... 91
7.4.6
Failure to Ramp ................................................................................................................................. 97
7.4.7
Dispatch Deviation in Non-Ramping Hours...................................................................................... 99
7.5
Profits or Benefits to the Market Participant ............................................................................................ 101
7.6
Expense or Disadvantage to the Market ................................................................................................... 103
7.7
Conclusion ............................................................................................................................................... 103
8.
ABITIBI’S CONDUCT IN RESPECT OF THE FORT FRANCES FACILITY ..................................... 105
8.1
CMSC Payments to Abitibi ...................................................................................................................... 105
8.2
Typical Operating Pattern ........................................................................................................................ 106
PUBLIC VERSION
4
8.2.1
Operating as a Net Load or Generator ............................................................................................. 106
8.2.2
Bidding Strategy .............................................................................................................................. 107
8.2.3
Ramping Pattern .............................................................................................................................. 107
8.2.4
Constrained-off CMSC Payments During Ramping ....................................................................... 109
8.2.5
Constrained-on CMSC Payments with Negative Bid Prices ........................................................... 113
8.2.6
Representative Pattern of Operation ................................................................................................ 117
8.3
Defects in Market Rules or Procedures .................................................................................................... 117
8.4
Exploitation of Constrained-Off CMSC................................................................................................... 118
8.4.1
Development of the Ramping CMSC Strategy ............................................................................... 120
8.4.2
Expanding the Magnitude of CMSC Using a High Bid Price ......................................................... 123
8.4.3
Ramping Faster than Submitted Ramp Rates .................................................................................. 134
8.4.4
Frequent Ramping ........................................................................................................................... 137
8.4.5
Combination of Load and Generator ............................................................................................... 143
8.4.6
Failure to Ramp ............................................................................................................................... 147
8.4.7
Dispatch Deviation in Non-Ramping Hours.................................................................................... 149
8.5
Exploitation of Constrained-On CMSC ................................................................................................... 151
8.5.1
Repayment of Portions of the Constrained-on CMSC..................................................................... 152
8.5.2
Development of the Negative-Price Bidding Strategy .................................................................... 154
8.5.3
Expanding the Magnitude of CMSC Using a Low Bid Price .......................................................... 161
8.5.4
Deviating from Dispatch Instructions to Appear to be Constrained On .......................................... 163
8.6
Profits or Benefits to the Market Participant ............................................................................................ 165
8.6.1
Constrained-off CMSC Payments ................................................................................................... 165
8.6.2
Constrained-on CMSC Payments .................................................................................................... 166
8.7
Expense or Disadvantage to the Market ................................................................................................... 168
8.8
Conclusion ............................................................................................................................................... 169
9.
REVIEW OF CONTINUING CMSC PAYMENTS AND RECENT DEVELOPMENTS REGARDING
REMEDIAL ACTION FOR GAMING .................................................................................. 171
9.1
Constrained-Off CMSC Payments ........................................................................................................... 171
9.2
Constrained-On CMSC Payments ............................................................................................................ 172
9.3
Continuing CMSC Payments ................................................................................................................... 172
9.4
Recent Developments Regarding Remedial Action for Gaming .............................................................. 174
10. SUMMARY OF FINDINGS AND RECOMMENDATION....................................................................... 176
10.1 Findings – Bowater .................................................................................................................................. 176
10.2 Findings – Abitibi .................................................................................................................................... 180
10.3 Recommendations .................................................................................................................................... 184
APPENDIX A
GLOSSARY ............................................................................................................................... 185
APPENDIX B
ABITIBI BOWATER INC. CORPORATE CHART ............................................................ 189
APPENDIX C
SELECTED BOWATER AND ABITIBI PERSONNEL WHO PREPARED OR
RECEIVED COMMUNICATIONS AND DOCUMENTS REFERRED TO IN THIS
REPORT .................................................................................................................................... 190
APPENDIX D
THE UNCONSTRAINED MODE AND SCHEDULE .......................................................... 192
APPENDIX E
THE CONSTRAINED MODE AND SCHEDULE ................................................................ 195
APPENDIX F
CALCULATION OF CONSTRAINED-OFF AND CONSTRAINED-ON CMSC
PAYMENTS .............................................................................................................................. 197
APPENDIX G
CMSC PAYMENTS ARISING FROM SELF-INDUCED RAMPING ............................... 199
APPENDIX H
IESO “BUSINESS RULES” FOR CLAWBACK OF CONSTRAINED-OFF CMSC
PAYMENTS .............................................................................................................................. 202
APPENDIX I
FIVE LARGEST CMSC PAYMENT DAYS FOR BOWATER’S THUNDER BAY
FACILITY ................................................................................................................................. 204
PUBLIC VERSION
5
APPENDIX J
ABITIBIBOWATER CANADA INC. PRESENTATION SLIDES DISCUSSING CMSC
AND OPERATING RESERVE ............................................................................................... 208
APPENDIX K
SAMPLE BOWATER CALCULATIONS FOR THE THUNDER BAY FACILITY
RELATING TO RAMPING DOWN FASTER THAN SUBMITTED RAMP RATES ..... 211
APPENDIX L
SAMPLE BOWATER ANALYSIS OF CMSC DURING RAMP DOWN .......................... 212
APPENDIX M
FIVE LARGEST CMSC PAYMENT DAYS FOR ABITIBI’S FORT FRANCES
FACILITY ................................................................................................................................. 213
APPENDIX N
RESOLUTE’S JULY 2, 2014 RESPONSE ............................................................................. 217
APPENDIX O
THE PANEL’S COMMENTS ON RESOLUTE’S JULY 2, 2014 RESPONSE AND
UPDATE ON SUBSEQUENT CORRESPONDENCE AND THE GENERAL CONDUCT
RULE ......................................................................................................................................... 247
PUBLIC VERSION
6
List of Tables
Table 4-1: Net Energy Cost per MWh for the Thunder Bay Facility, the Fort Frances Facility and All Other Loads
January – August 2010 ....................................................................................................................................... 21
Table 4-2: Energy Charges, CMSC Payments and Net Energy Cost for the Thunder Bay Facility, the Fort Frances
Facility and All Other Dispatchable Loads January – August 2010 .................................................................. 22
Table 6-1: Differences between the Unconstrained and Constrained Modes of the IESO’s Dispatch Algorithm ....... 33
Table 6-2: Simplified CMSC Formulas ....................................................................................................................... 35
Table 6-3: Illustration of CMSC Payments Arising from Differences in the Unconstrained and Constrained Modes of
the IESO’s Dispatch Algorithm ......................................................................................................................... 37
Table 7-1: Gross and Net CMSC Payments to Bowater for the Thunder Bay Facility February – August 2010 ........ 42
Table 7-2: Ramp Rates for the Thunder Bay Facility February – August 2010 .......................................................... 45
Table 7-3: Energy Charges and CMSC Payments on a Typical Ramp Down of the Thunder Bay Facility May 14,
2010, HE 6 ......................................................................................................................................................... 48
Table 7-4: Energy Charges and CMSC Payments on a Typical Ramp Up of the Thunder Bay Facility May 14, 2010,
HE 19 ................................................................................................................................................................. 49
Table 7-5: Estimated Impact of Bowater’s High Bid Prices on CMSC Payments for the Thunder Bay Facility
February – August 2010 ..................................................................................................................................... 72
Table 7-6: Likelihood of the Thunder Bay Facility Being Constrained Off at Various Bid Prices During Self-Induced
Ramping Hours January to December 2009 and February to August 2010 ....................................................... 77
Table 7-7: Estimated Impact of Bowater’s Ramp Down Timing on CMSC Payments for the Thunder Bay Facility
February – August 2010 ..................................................................................................................................... 90
Table 7-8: CMSC Payments on a Fast Ramp Down of the Thunder Bay Facility June 7, 2010, HE 6 ....................... 96
Table 8-1: Gross and Net CMSC Payments to Abitibi for the Fort Frances Facility January – August 2010 ........... 105
Table 8-2: Ramp Rates for the Dispatchable Load at the Fort Frances Facility January 2009 – August 2010.......... 108
Table 8-3: Energy Charges and CMSC Payments Received on a Typical Ramp Down of the Net Load at the
Fort Frances Facility March 21, 2010, HE 9 .................................................................................................... 111
Table 8-4: Energy Charges and CMSC Payments Received on a Typical Ramp Up of the Net Load at the
Fort Frances Facility March 21, 2010, HE 11 .................................................................................................. 113
Table 8-5: Consumption Bids for the Net Load at the Fort Frances Facility June 1, 2010 ........................................ 116
Table 8-6: Estimated Impact of Abitibi’s High Bid Prices on CMSC Payments for the Fort Frances Facility January
– August 2010 .................................................................................................................................................. 128
Table 8-7: Likelihood of the Fort Frances Facility Being Constrained-Off at Various Bid Prices During Self-Induced
Ramping Hours January 1, 2009 to August 28, 2010....................................................................................... 131
Table 8-8: CMSC Payments on a Fast Ramp Down of the Fort Frances Facility March 5, 2010, HE 3 ................... 136
Table 8-9: Self-Induced Ramps Greater than ● MW at the Fort Frances Facility 2007 to 2010 ............................... 142
PUBLIC VERSION
7
Table 9-1: Gross and Net CMSC for the Thunder Bay Facility, the Fort Frances Facility and All Other Dispatchable
Loads 2008 – 2013 ........................................................................................................................................... 173
Table 10-1: Summary of Findings Related to Bowater and the Thunder Bay Facility .............................................. 176
Table 10-2: Summary of Findings Related to Abitibi and the Fort Frances Facility ................................................. 180
Table I-1: Highest CMSC Daily Payments and Net Energy Cost for the Thunder Bay Facility January to August
2010.................................................................................................................................................................. 204
Table M-1: Highest CMSC Daily Payments and Net Energy Cost for the Fort Frances Facility January to August
2010.................................................................................................................................................................. 213
PUBLIC VERSION
8
List of Figures
Figure 7-1: Sample Ramping Pattern and CMSC Payments for the Thunder Bay Facility May 14, 2010 .................. 47
Figure 7-2: Excerpt from “Thunder Bay 2010 Power Cost – October 1st, 2009”, a presentation by Bowater to
AbitibiBowater Inc. ............................................................................................................................................ 57
Figure 7-3: Typical Ramp Down Pattern Used by the Thunder Bay Facility February – August 2010 ...................... 85
Figure 7-4: Alternative Ramp Down Bid Structure for the Thunder Bay Facility....................................................... 87
Figure 7-5: Scheduled and Actual Consumption and CMSC Payments at the Thunder Bay Facility During a Failure
to Ramp April 11, 2010 ...................................................................................................................................... 98
Figure 7-6: Scheduled and Actual Consumption and CMSC Payments During Non-Ramp Dispatch Deviation at the
Thunder Bay Facility July 10, 2010 ................................................................................................................. 100
Figure 8-1: Sample Ramping Pattern and CMSC Payments for the Net Load at the Fort Frances Facility March 21,
2010.................................................................................................................................................................. 110
Figure 8-2: Schedules and CMSC Payments During Hours with Negative Bid Prices for the Net Load at the Fort
Frances Facility June 1, 2010 ........................................................................................................................... 115
Figure 8-3: Schedules and CMSC Payments when Using the Generator to Ramp Changes in the Net Load at the
Fort Frances Facility July 28, 2010 .................................................................................................................. 144
Figure 8-4: Scheduled and Actual Consumption and CMSC Payments for the Net Load at the Fort Frances Facility
During a Failure to Ramp August 16, 2010 ..................................................................................................... 148
Figure 8-5: Schedules and CMSC Payments During Non-Ramp Dispatch Deviation at the Fort Frances Facility
March 14, 2010 ................................................................................................................................................ 150
Figure F-1: Illustration of How CMSC is Calculated Using the Maximum of AQEW and DQSW During a Ramp
Down ................................................................................................................................................................ 198
Figure G-1: Illustration of How CMSC May be Increased Using a Lower Ramp Rate During Self-Induced Ramp Up
.......................................................................................................................................................................... 200
Figure G-2: Illustration of How CMSC May be Increased Using a Lower Ramp Rate During Self-Induced Ramp
Down ................................................................................................................................................................ 201
Figure I-1: Schedules and CMSC Payments for the Thunder Bay Facility April 11, 2010 ....................................... 205
Figure I-2: Schedules and CMSC Payments for the Thunder Bay Facility March 5, 2010 ....................................... 205
Figure I-3: Schedules and CMSC Payments for the Thunder Bay Facility February 21, 2010 ................................. 206
Figure I-4: Schedules and CMSC Payments for the Thunder Bay Facility May 16, 2010 ........................................ 206
Figure I-5: Schedules and CMSC Payments for the Thunder Bay Facility April 26, 2010 ....................................... 207
Figure M-1: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility June 1, 2010 ............... 214
Figure M-2: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility July 21, 2010 .............. 214
Figure M-3: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility July 22, 2010 .............. 215
Figure M-4: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility April 16, 2010 ............ 215
Figure M-5: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility June 2, 2010 ............... 216
PUBLIC VERSION
9
REPORT ON AN INVESTIGATION INTO POSSIBLE GAMING BEHAVIOUR
RELATED TO CONGESTION MANAGEMENT SETTLEMENT CREDIT PAYMENTS
BY ABITIBI-CONSOLIDATED COMPANY OF CANADA AND BOWATER
CANADIAN FOREST PRODUCTS INC.
____________________________________________________________
1.
EXECUTIVE SUMMARY
This Report sets out the findings of the Market Surveillance Panel (the “Panel”) in relation to an
investigation into Congestion Management Settlement Credit (“CMSC”) payments received by
Bowater Canadian Forest Products Inc. (“Bowater”) and Abitibi-Consolidated Company of
Canada (“Abitibi”) during the eight-month period from January 2010 to August 2010.1 The
companies were both ultimately owned by Abitibi Bowater Inc. (renamed Resolute Forest
Products Inc. in 2011).
The Panel is mandated to monitor and investigate activities in the wholesale electricity market
and the conduct of market participants, including in relation to inappropriate or anomalous
market conduct. The conduct that is the subject of this investigation was noted by the Panel in
one of its semi-annual Monitoring Reports, which discussed anomalous CMSC payments being
made to two dispatchable loads located in Northern Ontario. Following receipt and publication
of the Panel’s Monitoring Report, the then Chair of the Ontario Energy Board (“OEB”)
requested that the Panel investigate the matter. At the same time, the Independent Electricity
System Operator (“IESO”) moved expeditiously to deal with two of the major sources of CMSC
payments that had been made to Abitibi and Bowater and that are described in this Report.
As set out in greater detail below, the Panel’s investigation considered a number of aspects of
Bowater’s and Abitibi’s market conduct, including the submission of extremely high bid prices
(both market participants), the submission bid quantities above the level of electricity that the
facility was generally capable of consuming (Bowater) and frequent ramping (Abitibi). These
kinds of behaviours can be used to obtain CMSC payments from the wholesale market in a
manner and in amounts that go beyond what is intended by the wholesale market design and the
1
Capitalized terms and abbreviations are listed and defined in the Glossary (Appendix A).
PUBLIC VERSION
Report on an Investigation into Possible Gaming Behaviour Related to Congestion Management
Settlement Credit Payments by Abitibi-Consolidated Company of Canada and Bowater Canadian Forest
Products Inc.
10
rules that govern the markets.2 Where a market participant exploits a defect in the design, rules
or procedures governing the wholesale electricity markets, and obtains a profit or benefit at the
expense or disadvantage of the market, the Panel considers that to be gaming.
The Panel has concluded that both companies engaged in gaming. By their market conduct, they
exploited certain market defects and, in so doing, received $20.4 million in CMSC payments
during the eight-month period in question, and there was a corresponding disadvantage or
expense to the market.3 The documents and materials obtained by the Panel for the purposes of
this investigation reveal that Bowater’s and Abitibi’s conduct was deliberate, and was
understood by the companies to be inconsistent with the principles underlying the CMSC
framework and as having the potential to be regarded as gaming.
1.1
CMSC Payments
To alleviate congestion resulting from transmission system constraints, the IESO must
sometimes instruct a dispatchable load4 to consume more or less energy than the load had bid to
consume. When a load is “constrained off”, it is being instructed to consume less energy than it
desires even though the load’s bid price is higher than the prevailing market price. When a load
is “constrained on”, it is being instructed to consume more energy than it desires at a time when
the prevailing market price is higher than the load’s bid price. In either case, the dispatchable
load’s operating profit would be assumed to be reduced.
CMSC payments are intended to compensate a load for the assumed reduction in operating profit
caused by following such IESO dispatch instructions. Although the rationale for CMSC
payments is clear, the rules and procedures governing the calculation of the payments are
complex. The Panel identified various defects in the Market Rules and IESO procedures (as they
2
See section 6.3 for further detail regarding the origin, purpose and calculation of CMSC payments.
The total CMSC payments made to the two companies over the eight-month period was over $22 million. $20.4
million is the amount that the Panel has found to have been received as a result of gaming.
4
Bowater and Abitibi were registered as “dispatchable loads” in the wholesale electricity market. They submitted
bids (the prices and quantities of electricity they were willing to purchase) into the market every hour and would be
scheduled by the IESO as long as their bid prices were not lower than the market price. Dispatchable loads are
required to adhere to IESO dispatch instructions sent every five minutes that indicate the amount of power they
should consume.
3
PUBLIC VERSION
Products Inc.
11
existed at the relevant time) which Bowater and Abitibi exploited to obtain CMSC payments that
were self-induced, rather than being caused by conditions on Ontario’s power grid. “Selfinduced” refers to the ability of a market participant to bring about an outcome (i.e., CMSC
payment, ramp or dispatch instruction) through its own actions.
CMSC payments made by the IESO are recovered from market participants based on their
respective withdrawals from the IESO-controlled grid through what is known as an “Uplift”
charge. Ultimately, the cost of CMSC payments is borne by all electricity consumers.
1.2
Bowater’s Conduct
Most of the CMSC payments received by Bowater were triggered in hours when its Thunder Bay
pulp and paper mill was voluntarily reducing (“ramping down”) or increasing (“ramping up”) its
power consumption. The CMSC payments received by Bowater during self-induced ramping
hours far exceeded the cost of the electricity it consumed during those hours. As a result, it was
effectively being paid, rather than paying, to consume the amount of electricity it wanted to
consume during such hours.
The main behaviours that led to Bowater’s substantial CMSC payments were:
·
Submitting an extremely high bid price during the hours that Bowater chose to ramp
its facility up or down.
·
Submitting bid quantities above the level of electricity that its facility was generally
capable of consuming.
·
Timing the ramping down of its facility to increase the amount of CMSC payments.
·
Submitting ramp rates that understated the rate at which its facility changed its
electricity consumption, thereby increasing its CMSC payments.
Bowater received a total of $12.3 million in CMSC payments in the eight-month period in
question. The Panel determined that the overwhelming majority of those payments – $11.0
million – was triggered by Bowater’s gaming behaviour, which in turn increased Uplift charges
PUBLIC VERSION
Products Inc.
12
for all wholesale customers by $0.12/MWh during that eight-month period. The $11.0 million in
CMSC payments also served to effectively reduce Bowater’s net cost for electricity at its
Thunder Bay pulp and paper mill to an amount well below the net energy cost of other
dispatchable wholesale customers.5
In summary, the Panel concluded that the four above-noted behaviours exploited market defects
in the CMSC regime, were highly profitable to Bowater and disadvantaged the market
participants who pay Uplift charges. These behaviours therefore constituted gaming.
1.3
Abitibi’s Conduct
Abitibi received significant CMSC payments during hours when it was voluntarily ramping its
Fort Frances pulp and paper mill up or down. The CMSC payments received by Abitibi during
self-induced ramping hours far exceeded the cost of the electricity it consumed. As a result, it
was effectively being paid, rather than paying, to consume the amount of electricity it wanted to
consume during such hours.
The main behaviours engaged in by Abitibi that led to substantial constrained-off CMSC
payments were:
·
Submitting an extremely high bid price during the hours that Abitibi chose to
ramp its facility up or down.
·
Submitting ramp rates that understated the rate at which its facility changed its
electricity consumption, thereby increasing its CMSC payments.
·
Frequent ramping of the facility.
Between April and August 2010, Abitibi implemented an additional strategy to obtain
constrained-on CMSC payments during certain hours. It submitted an extremely negative bid
5
See Table 4-1 for the net energy cost for the Thunder Bay Facility and all other loads during the eight-month
period in question.
PUBLIC VERSION
Products Inc.
13
price and then either was constrained on or consumed above the level of its dispatch instructions.
A negative bid price should mean that a load is only willing to consume electricity if it is paid to
do so. However, Abitibi intended to and did consume electricity during many of these hours.
This behaviour relates to a market defect that had been publicly identified as such by the Panel
and the IESO at the time, and that the IESO had announced would be the subject of Market Rule
amendments.
Abitibi received a total of $9.7 million in net CMSC payments in the eight-month period in
question. The Panel determined that the overwhelming majority of those payments – $9.4
million – were triggered by Abitibi’s gaming behaviour, which in turn increased Uplift charges
for all wholesale customers by $0.09/MWh during that eight-month period. The $9.4 million in
CMSC payments also served to effectively reduce Abitibi’s net cost for electricity at its Fort
Frances pulp and paper mill to the point where Abitibi was in fact being paid to consume
electricity.6
In summary, Abitibi engaged in three behaviours to exploit market defects in the constrained-off
CMSC regime and two behaviours to exploit market defects in the constrained-on CMSC
regime. These behaviours were highly profitable to Abitibi and disadvantaged the market
participants who pay Uplift charges. These behaviours therefore constituted gaming.
1.4
Observations regarding Remedial Action and Review of Continuing CMSC
Payments
1.4.1 Remedial Action
The Panel encourages the IESO to take whatever action may be open to it to recover the amounts
paid to Bowater and Abitibi as a result of conduct that the Panel has found to constitute gaming
behaviour.
The Panel’s responsibilities include monitoring, investigations and reporting in respect of the
wholesale market. The Panel’s investigation reports may include recommendations, including
6
See Table 4-1 for the net energy cost for the Fort Frances Facility and all other loads during the eight-month period
in question.
PUBLIC VERSION
Products Inc.
14
recommendations regarding Market Rule amendments. However, the Panel does not have the
legislative mandate to impose sanctions or remedies when it finds that gaming has occurred.
While a compliance and enforcement regime exists in relation to breaches of the Market Rules,
gaming does not necessarily constitute a breach of the Market Rules.
The IESO is currently engaging in stakeholder consultations regarding the introduction of a
“general conduct rule” to the Market Rules. The Panel supports this initiative, and encourages the
IESO to proceed expeditiously with its consultations and to ensure that any rule that it
implements captures the conduct that is the subject of this investigation or similar kinds of
conduct that have been discussed in other Panel reports.
1.4.2 Review of Continuing CMSC Payments
In late August 2010, the IESO used an Urgent Market Rule Amendment to suspend all CMSC
payments to constrained-off dispatchable loads in light of the fact that significant CMSC
payments had been made to two dispatchable loads which the IESO believed to be inconsistent
with the intent of the CMSC regime. This foreclosed any further such payments to Bowater and
Abitibi. After stakeholder consultations, the IESO went on to implement two amendments to the
Market Rules in late 2010: one largely and permanently eliminated deviation-induced
constrained-on CMSC payments and the other permanently eliminated constrained-off CMSC
payments for self-induced ramping by dispatchable loads. These amendments dealt with two of
the major sources of CMSC payments to Bowater and Abitibi that are the subject of this Report.
However, dispatchable loads continue to receive CMSC payments. During 2011 to 2013,
Bowater received approximately $1.7 million in CMSC payments, Abitibi received $2.4 million,
and other dispatchable loads received $23.7 million. On December 10, 2012, Bowater generally
stopped bidding as a dispatchable load, and on September 12, 2013 Abitibi did the same. Both
market participants were therefore ineligible for CMSC payments. However, in 2013 CMSC
payments to Abitibi (prior to September 12) and other dispatchable loads remained significant at
$1.0 million and $13.3 million respectively.
The Panel considers that the continuing magnitude of CMSC payments under the current Market
Rules is significant enough to warrant further review. The Panel therefore recommends:
PUBLIC VERSION
Products Inc.
15
a)
The IESO should review the CMSC payments being made to dispatchable
loads since the November/December 2010 amendments to the Market
Rules in order to determine whether there are significant amounts that
continue to be unwarranted (i.e., paid as a result of market participant
actions rather than to compensate for operating profit reductions arising
from responding to dispatch instructions caused by Grid Conditions).
b)
If necessary, the IESO should make further amendments to the Market
Rules to eliminate unwarranted CMSC payments to dispatchable loads.
1.5
Postscript
In accordance with section 7.2.2 of the Ontario Energy Board’s By-law No. 3, the Panel
provided a draft of this Report to the market participants on April 16, 2014, to provide them with
an opportunity to discuss the findings with the Panel, to respond to the findings and to comment
on matters of factual accuracy and confidentiality. The Panel offered to meet with the market
participants, and identified the date by which any written response should be provided.
On May 15, 2014, Resolute FP Canada Inc. (“Resolute”), successor in interest to Abitibi and
Bowater, requested that the Panel provide data used to support the Panel’s findings before
responding to those findings. On June 6, 2014, the Panel provided Resolute with a large amount
of interval-by-interval data for each of the Thunder Bay and Fort Frances Facilities for the period
covered by the Panel’s investigation, being IESO data that the Panel used in its
analysis. Resolute delivered a written response to the Panel’s draft report on July 2, 2014.
Resolute’s July 2, 2014 response is reproduced in Appendix N, and the Panel’s comments on that
response are set out in Appendix O. Appendix O also describes subsequent correspondence
exchanged between Resolute and the Panel, as well as an update on the status of the IESO’s
“general conduct rule” referred to in section 1.4.1 above.
In terms of confidentiality, as part of its July 2, 2014 response Resolute requested that the
following be redacted from the public version of this Report: (i) numbers and figures such as bid
numbers and operating costs; and (ii) the names of Resolute personnel. Although the Panel
PUBLIC VERSION
Products Inc.
16
questions the commercially sensitive nature of the data referred to in (i) given the change in
status of the two facilities at issue, the Panel nonetheless agreed to redact certain data as well as
the names and titles of Resolute personnel. In accordance with section 7.5 of the Ontario Energy
Board’s By-law No. 3, both public and confidential versions of this Report have therefore been
prepared, the former for public communication and the latter for transmittal to the Chair of the
Ontario Energy Board and the CEO of the IESO. Appendix N as it appears in the public version
of this Report was redacted by Resolute.
With the exception of this Postscript, Appendices N and O, and section 7.4.3 (which was
modified by the Panel in light of Resolute’s response), this Report is as at December 31, 2013.
PUBLIC VERSION
Products Inc.
17
2.
INTRODUCTION
This Report contains the analysis and findings of the Market Surveillance Panel (“MSP”, or the
“Panel”) in respect of an investigation (the “Investigation”) into possible gaming of Congestion
Management Settlement Credit (“CMSC”) payments by dispatchable loads7 operated by
Abitibi-Consolidated Company of Canada (“Abitibi”) and Bowater Canadian Forest Products
Inc. (“Bowater”) during the period January 2010 to August 2010 (the “Relevant Period”).
Abitibi and Bowater are both subsidiaries of Abitibi Bowater Inc., and the issues relating to each
participant are being dealt with in a single report because of overlaps in some of the time periods,
personnel and behaviour related to both loads.
This Report begins by describing the market participants that are the subject of the Investigation
(Section 3) and the CMSC payments they received (Section 4). It also summarizes the Panel’s
investigation framework and process as well as the applicable Market Rules and Independent
Electricity System Operator (“IESO”) procedures (Section 5), and other relevant aspects of the
design of the Ontario wholesale market (Section 6). It then provides the Panel’s analysis,
findings and recommendations in respect of Bowater’s (Section 7) and Abitibi’s (Section 8)
activities. The Report concludes with a recommendation relating to possible continuing
unwarranted CMSC payments (Section 9).
With the exception of the Postscript that appears in section 1.5 of the Executive Summary and
the section and Appendices noted at the end of that Postscript, the information set out in this
Report is as at December 31, 2013.
7
Dispatchable loads are large industrial users of electricity that receive instructions from the electricity system
operator indicating the amount of electricity they should consume.
PUBLIC VERSION
Settlement Credits by Abitibi-Consolidated Company of Canada and Bowater Canadian Forest Products
Inc.
18
3.
THE MARKET PARTICIPANTS AND DISPATCHABLE FACILITIES
The dispatchable loads that are the subject of this investigation were owned during the relevant
period by subsidiaries of Abitibi Bowater Inc. (“ABI”), a Delaware registered corporation.8 In
2009, ABI and its Canadian and U.S. subsidiaries entered into bankruptcy proceedings in Canada
and the United States. ABI and its subsidiaries completed a reorganization and emerged from
creditor protection under the Companies' Creditors Arrangement Act (“CCAA”)9 in Canada and
under comparable US legislation on December 9, 2010.10 ABI has been renamed Resolute Forest
Products Inc.11
Although the two dispatchable facilities were under common ownership (see the corporate chart
in Appendix B), they were owned by different subsidiaries and each was registered as a separate
market participant with the IESO.
3.1
Bowater Canadian Forest Products Inc.
During the Relevant Period, Bowater owned and operated a pulp and paper mill in Thunder Bay,
Ontario. Bowater was owned by AbitibiBowater Canada Inc., a TSX-listed corporation.
AbitibiBowater Canada Inc. was owned by Bowater Canadian Holdings Incorporated, which was
owned by Bowater Incorporated (a Delaware corporation), which was owned by ABI. 12
Bowater’s facility in Thunder Bay (the “Thunder Bay Facility”) includes a thermo-mechanical
pulpmill (“TMP”) with two mainline refiners, a rejects refiner, auxiliaries, and a recycle mill.
The Thunder Bay Facility produced commercial printing papers, newsprint and market pulp. It
had a maximum dispatch capability of ● - ● MW and was one of the largest dispatchable loads in
8
See Abitibi Bowater Inc. Corporate Chart in Appendix B.
Companies’ Creditors Arrangement Act, 1985 (Canada), as amended, online: http://lawslois.justice.gc.ca/eng/acts/C-36/page-1.html
10
Resolute Forest Products Inc., webpage, “About Us: Emergence”, as at November 5, 2012, online:
https://web.archive.org/web/20121105050415/http://www.resolutefp.com/emergence/.
11
Resolute Forest Products Inc., webpage, “About Us: Legal Entity Name Changes”, as at February 26, 2014,
online: http://www.resolutefp.com/About_Us/Identity/
12
Responses to RFI (defined in Section 5.4), A.1, p. 1.
9
PUBLIC VERSION
Products Inc.
19
Ontario. During the Relevant Period, Bowater held an Electricity Wholesaler Licence from the
Ontario Energy Board (“OEB”) which authorized it to buy and sell electricity through the IESOadministered markets.13
From July 2003 to August 2006, the Thunder Bay Facility operated as a dispatchable load. It
elected to become non-dispatchable in September 2006. In 2009, as part of the CCAA process,
Bowater idled its machines and reassessed its operations. It addressed labour, power and fibre
cost issues, and also consulted with personnel at affiliated entities regarding opportunities to
reduce power costs by generating CMSC payments as a dispatchable load.14 The Thunder Bay
Facility resumed operation of its Paper Machine 5 on December 17, 2009.15 It became a
dispatchable load again in the IESO-administered market on February 8, 2010.
3.2
Abitibi-Consolidated Company of Canada
During the Relevant Period, Abitibi16 owned and operated a pulp and paper mill in Fort Frances,
Ontario. Abitibi was owned by Abitibi-Consolidated Inc., which was owned by Abitibi Bowater
Canada Inc. (18%), a Canadian corporation, and ABI (82%). Abitibi Bowater Canada Inc. was
also ultimately owned by ABI.
Abitibi’s facility in Fort Frances (the “Fort Frances Facility”) includes three paper machines
(two of which were active), one kraft mill, and a biomass and natural gas boiler/generator.17 The
facility produced commercial printing papers and market pulp. The load had a maximum
dispatch capability of ● MW and the generator had a maximum dispatch capacity of ● MW.
13
Responses to RFI, A.4, p. 113; Electricity Wholesaler Licence EW-2005-0537 Bowater Canadian Forest Products
Inc., available online at:
http://www.rds.ontarioenergyboard.ca/webdrawer/webdrawer.dll/webdrawer/rec/43273/view/DO_Bowater_licence_
20060125.PDF.
14
See, e.g., email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], September 11, 2009.
Responses to RFI, B.2.5.
15
Responses to RFI, B.2, p.1.
16
Abitibi-Consolidated Company of Canada was continued as Abibow Canada Inc. on December 10, 2010; Abibow
Canada Inc. subsequently changed its name to Resolute FP Canada Inc..
17
PUBLIC VERSION
Products Inc.
20
During the Relevant Period, Abitibi held an Electricity Generation Licence from the OEB which
authorized it to generate, sell and buy electricity through the IESO-administered markets.18
The Fort Frances Facility has operated as a dispatchable load since September 2004 and as a
dispatchable generator since June 2007. During the CCAA restructuring, the Fort Frances
Facility shut down one of the paper machines. At times, this resulted in changes to its operating
pattern because of limited pulp storage capacity. On January 22, 2010, Fort Frances began to
operate as an aggregated facility, meaning the load and generator could effectively bid or offer,
and pay or be paid, for electricity as either a net load or net generator.19 It typically operated as a
net load (i.e. consumption exceeding on-site generation) and was charged for energy on the basis
of its net metered consumption.
18
Responses to RFI, A.4, p. 106; Electricity Generation Licence EG-2003-0204 Abitibi-Consolidated Company of
Canada. The amended licence is available online at:
http://www.rds.ontarioenergyboard.ca/webdrawer/webdrawer.dll/webdrawer/rec/384821/view/amd_licence_eg_reso
lute_20130228.PDF.
19
Pursuant to Chapter 7, Section 2.3 of the Market Rules.
PUBLIC VERSION
Inc.
21
4.
CMSC PAYMENTS TO BOWATER AND ABITIBI
Table 4-1 summarizes the CMSC payments received relative to the cost of purchasing electricity
for each of the two facilities and all other loads in Ontario during the Relevant Period. CMSC
payments reduced Bowater’s average net electricity cost to $25.44/MWh, more than 60% less
than the price being paid by other loads. Abitibi received approximately $2.5 million in net
energy payments (i.e. its CMSC payments exceeded the energy, Uplift and Global Adjustment
charges) and it was effectively being paid to consume energy at an average rate of $22.47/MWh.
Table 4-1: Net Energy Cost per MWh for the Thunder Bay Facility,
the Fort Frances Facility and All Other Loads
January – August 2010
($000, MWh and $/MWh)
Participant
Energy
Charges
($000)*
CMSC
($000)
**
Net
Energy
Cost
($000)
8,587
(2,518)
155,325
5,946,423
Consumption
(MWh)
Net Energy
Cost
Per MWh
($/MWh)
25.44
(22.47)
64.00
67.03
Bowater (Thunder Bay Facility)***
20,921
12,334
337,514
Abitibi (Fort Frances Facility)
7,176
9,694
112,034
All Other Dispatchable Loads
156,413
1,088
2,427,092
Non-Dispatchable Loads
5,946,423
n/a
88,709,646
* Includes Global Adjustment and Uplift charges.
** All amounts are net CMSC after clawback of charge type 105 (CMSC paid for the difference between the
constrained and unconstrained schedule) and charge type 1050 (CMSC that should not be paid because it was the
result of that registered facility’s own equipment or operational limitations according to IESO Business Rules). Also
excludes voluntary repayments. The clawback adjustments and the IESO Business Rules are described in Section
6.3.4 and Appendix H.
*** Bowater data excludes January 2010 as it only resumed being a dispatchable load in February 2010.
Table 4-2 summarizes the monthly CMSC payments made to the two facilities and to all other
dispatchable loads during the Relevant Period. It also provides comparative data about the cost
of electricity consumed (including Uplift and Global Adjustment charges). The CMSC amounts
are net payments (i.e. gross CMSC less any clawbacks and voluntary repayments). Over the
eight-month period the Thunder Bay Facility, which represented 10% of dispatchable load
capacity in Ontario, received 53% of CMSC payments made to all Ontario dispatchable loads.
The Fort Frances Facility, which represented 7% of dispatchable load capacity in Ontario,
received 42% of such payments.
PUBLIC VERSION
Products Inc.
22
Table 4-2: Energy Charges, CMSC Payments and Net Energy Cost for the Thunder Bay
Facility, the Fort Frances Facility and All Other Dispatchable Loads
($000)
Month
(2010)
Bowater
Abitibi
(Thunder Bay Facility)
(Fort Frances Facility)
Energy
Charges
*
CMSC
**
Net
Energy
Cost
Energy
Charges
*
CMSC
**
Net
Energy
Cost
All Other
Dispatchable Loads
Energy
Charges
*
CMSC
**
Net
Energy
Cost
Jan
Not Dispatchable
1,875
385
1,490
21,232
64
21,296
Feb
2,859
1,405
1,454
448
413
35
19,154
130
19,284
Mar
3,538
2,508
1,030
608
1,134
(526)
23,884
156
24,041
Apr
3,094
2,602
493
840
1,857
(1,017)
23,450
102
23,553
May
3,124
2,339
785
1,136
1,280
(144)
19,343
178
19,521
Jun
2,936
1,796
1,141
681
2,689
(2,009)
16,915
249
17,164
Jul
2,676
867
1,809
816
1,644
(827)
16,628
153
16,781
Aug
2,693
818
1,875
773
292
481
15,806
56
15,862
Total
$20,921 $12,334
$8,587
$7,176 $9,694
($2,518)
$156,413 $1,088
$157,501
** All amounts are net CMSC after clawback of charge type 105 (CMSC paid for the difference between the
constrained and unconstrained schedule) and charge type 1050 (CMSC that should not be paid because it was the
result of that registered facility’s own equipment or operational limitations according to IESO Business Rules). Also
excludes voluntary repayments. The clawback adjustments and the IESO Business Rules are described in Section
6.3.4 and Appendix H.
PUBLIC VERSION
Products Inc.
23
5.
INVESTIGATION PROCESS AND FRAMEWORK
This section provides an overview of the MSP mandate in respect of market monitoring and
gaming investigations, background on the events leading to the request for and commencement
of the Investigation, the information gathered and the analytical framework used to assess
gaming.
5.1
Market Surveillance Panel Mandate
The MSP is empowered under the Electricity Act, 1998 (the “Act”) to conduct investigations into
any activity related to the IESO-administered markets or the conduct of a market participant.20
The MSP, with the support of the IESO’s Market Assessment Unit (“MAU”),21 is also required
by OEB By-Law #3 (the “MSP By-Law”) to monitor activities related to the IESO-administered
markets and the conduct of market participants with a view to identifying, among other matters:
·
inappropriate or anomalous market conduct, including possible abuses of market power
and gaming;
·
design flaws and inefficiencies in the Market Rules and other rules and procedures of the
IESO; and
·
design flaws in the overall structure of the IESO-administered markets.22
The general process applicable to MSP investigations is set out in the MSP By-Law which
provides, among other things, that:
·
the Panel may initiate an investigation on its own, upon receipt of a complaint or at the
request of the OEB Chair;23
20
Electricity Act, 1998 (Ontario), as amended, online: http://www.elaws.gov.on.ca/html/statutes/english/elaws_statutes_98e15_e.htm#BK95,s.37(1).
21
The MAU provides support to the MSP pursuant to the Protocol Related to Market Surveillance Panel
(“Protocol”) between the IESO and the OEB, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/msp_protocol.pdf. References in this report to
investigative steps carried out by the Panel include investigative steps carried out by the MAU on behalf of the
Panel.
22
MSP By-Law, as amended, online:
http://www.ontarioenergyboard.ca/oeb/_Documents/About%20the%20OEB/OEB_bylaw_3.pdf, s. 4.1.1.
23
Ibid, s. 5.1.1.
PUBLIC VERSION
Products Inc.
24
·
where the Panel commences an investigation, the Panel shall, upon determining that there
is a prima facie case in respect of the conduct of a person that is the subject matter of the
investigation, notify that person of the commencement of the investigation;24
·
for the purpose of carrying out an investigation, the Panel has the power to examine and
compel the production of any documents or other things, to summon and compel
testimony, to conduct inspections, and to obtain warrants for search and seizure as
authorized by the Act;25 and
·
upon completion of an investigation, the Panel shall prepare a written report on the matter
investigated, the Panel’s findings and its recommendations, if any. 26
5.2
Background to Investigation
The Investigation arose as a result of market monitoring activities conducted by the MAU on
behalf of the Panel during the spring and summer of 2010.
5.2.1 Initial Inquiries by the MAU
In May 2010, the MAU observed that large amounts of CMSC payments were being made to
Bowater’s Thunder Bay Facility and Abitibi’s Fort Frances Facility. It appeared that
anomalously high CMSC payments were arising as a result of facility-specific behaviours at both
Facilities. The MAU briefed the Panel, and the Panel asked the MAU to examine the issues in
greater detail.
On June 11, 2010, MAU staff contacted Bowater and Abitibi personnel to discuss the CMSC
payments that the two Facilities were receiving. Bowater and Abitibi were subsequently
provided with a summary which identified three high-level factors that, in the MAU’s view,
appeared to be contributing to the CMSC payments: (i) ramping actions; (ii) deviation from
intended consumption in the market schedule resulting in “constrained-on” CMSC payments at
the Fort Frances Facility; and (iii) deviation from intended consumption in the market schedule
24
Ibid, s. 5.1.9.
Ibid, s. 5.1.11.
26
Ibid, ss. 5.1.13 and 7.2.
25
PUBLIC VERSION
Products Inc.
25
resulting in “constrained-off” payments.27 During discussions with a representative of the Fort
Frances and Thunder Bay Facilities in June 2010, the MAU also referenced information
regarding the potential scope of gaming activity contained in the Panel’s Monitoring Document:
Monitoring of Offers and Bids in the IESO-Administered Electricity Markets.28
The MAU noted that the Thunder Bay Facility was often receiving upward of $80,000 in CMSC
payments per day compared to the approximately $8,000 in CMSC payments per day it received
in 2006 (when the facility had previously been dispatchable). It also indicated that Bowater’s
$●/MWh bid price was contributing to the very large CMSC payments (the relationship between
bid price and CMSC payments is explained in Section 6.3.2). As it has done on various
occasions in the past with market participants, the MAU requested on behalf of the IESO that
Bowater consider a voluntary repayment of all CMSC payments associated with ramping
(Bowater’s ramping pattern is described in Section 7.2.2), and noted that a bid price of
$●/MWh29 would be similar to offer price changes that had been adopted by various generators
who receive CMSC payments when they voluntarily chose to ramp down.30 Bowater declined to
repay any ramping CMSC payments31, although it lowered its bid price from $●/MWh to
$●/MWh during ramping periods on a go-forward basis.32
The MAU also noted the numerous instances of constrained-on CMSC payments at the Fort
Frances Facility which appeared to be self-induced. The MAU requested that Abitibi consider a
27
Email from MAU to [Senior Abitibi Personnel #2] dated June 24, 2010. Responses to RFI, B.13.27.
MSP, Monitoring Document: Monitoring of Offers and Bids in the IESO-Administered Electricity Markets, March
3, 2010, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MSP_Monitoring_Offers_Bids_Document_20100310.
pdf, p. 44.
29
Email from MAU to [Senior Abitibi Personnel #2], June 24, 2010. Responses to RFI, B.13.27.
30
For further information about the relationship between CMSC payments and offer prices used by generators to
signal an intention to ramp down, see MSP, Monitoring Document: Generator Offer Prices Used to Signal an
Intention to Come Offline, August 19, 2011, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MonitoringDocument_GeneratorOfferPrices_20110819.
pdf; and MSP, Monitoring Report on the IESO-Administered Electricity Markets for the period from May 2008 –
October 2008, online: http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/msp_report_200901.pdf, p. 213.
31
Email from [Senior Abitibi Personnel #2] to MAU, June 25, 2010. Responses to RFI, B.13.27. See also email
from [Senior Abitibi Personnel #2] to MAU, August 16, 2010. Responses to RFI, B.13.88
32
Emails from [Senior Abitibi Personnel #2] to MAU, June 30, 2010. Responses to RFI, B.13.40.
28
PUBLIC VERSION
Products Inc.
26
voluntary repayment of ramping CMSC and constrained-on CMSC payments.33 Abitibi has not
repaid any ramping CMSC. However, it repaid the portion of constrained-on CMSC which arose
as a result of consumption deviation between April 2010 and July 2010, which it calculated to be
$1.825 million.34
5.2.2 MSP Winter 2010 Monitoring Report
On August 30, 2010, the MSP published its semi-annual Monitoring Report for the period from
November 2009 to April 2010 (the “Winter 2010 Monitoring Report”) which contained
information about the anomalous CMSC payments made to two dispatchable loads located in
Northwestern Ontario.35 Bowater and Abitibi were not named in the Monitoring Report.
5.3
Request for an Investigation
Following receipt and publication of the Winter 2010 Monitoring Report, the then Chair of the
OEB wrote to the then Chair of the MSP on September 3, 2010 and requested that the Panel
investigate the circumstances that lead to the anomalous CMSC payments being made to two
dispatchable loads.36 The Panel commenced the Investigation in response to the OEB Chair’s
request and notified Bowater and Abitibi that a gaming investigation had been commenced.
5.4
Information Gathering
In carrying out its Investigation, the Panel obtained and considered extensive information from
the IESO. This included statistical information related to prices, scheduled and actual
consumption, settlement payments and other data.
The Panel also requested extensive information from Bowater and Abitibi. Information and
materials were provided by Bowater and Abitibi in response to the Panel’s requests for
information (the “Responses to RFI”) without the Panel having to use its statutory inspection or
33
Email from MAU to [Senior Abitibi Personnel #2], June 24, 2010. Responses to RFI, B.13.27.
Email from [Senior Abitibi Personnel #2] to MAU, August 17, 2010. Responses to RFI, B.13.163.
35
MSP, Monitoring Report on the IESO-Administered Electricity Markets for the period from November 2009 –
April 2010, online: http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MSP_Report_20100830.pdf. p. 112.
36
The request was made pursuant to MSP By-Law, s. 5.1.1(c).
34
PUBLIC VERSION
Products Inc.
27
other compulsory powers. Bowater and Abitibi represented that they had provided correct and
complete responses to the Panel’s information requests. The information provided by Bowater
and Abitibi included:
·
copies of emails between personnel at the companies and their affiliates37 that pertained
directly or indirectly to CMSC payments received during the Relevant Period;
·
copies of communications and documents related to the development of bidding
strategies in the wholesale electricity market and associated financial implications; and
·
copies of documents related to operating strategies, including the determination of ramp
rates and consumption patterns.
The Panel conducted detailed assessments of Bowater’s and Abitibi’s market conduct, their
Responses to RFIs and relevant market outcomes. The assessments were based on the analytical
framework used to assess gaming issues.
5.5
Framework for Gaming Investigations
This section outlines the framework applied by the Panel to assess whether the behaviours of a
market participant constitute gaming.
The Panel’s mandate includes investigations in relation to conduct that may constitute an
abuse of market power or gaming. In the course of providing a framework for analyzing
market power issues, the Panel has noted that gaming is a separate concept (which may or
may not overlap with market power concerns) that encompasses, among others, market
manipulation and conduct that involves the following four elements:
(i)
a defect in the market design, poorly specified rules or procedures or a gap
in the Market Rules or procedures (collectively referred to as a “market
defect”);
37
See Appendix C for a list of selected Bowater, Abitibi and affiliated company personnel who prepared or received
the communications and documents referenced in this Report.
PUBLIC VERSION
Products Inc.
28
(ii)
exploitation of the market defect by the market participant;
(iii)
profit or other benefit to the market participant; and
(iv)
expense or disadvantage to the market. 38
Sections 7 and 8 of this Report address each of these elements in respect of the conduct of
Bowater at the Thunder Bay Facility and Abitibi at the Fort Frances Facility, respectively.
Section 6 provides contextual information about relevant aspects of the wholesale market.
38
See Market Surveillance Panel, Report on an Investigation into Possible Gaming Behaviour Related to Infeasible
Import Transactions by TransAlta Energy Marketing Corp. on the Manitoba-Ontario Intertie, Investigation
No. 2011-02, October 22, 2012, p. 7.
PUBLIC VERSION
Products Inc.
29
6.
RELEVANT ASPECTS OF THE WHOLESALE MARKET DESIGN
This section provides an overview of how dispatchable loads participate in the Ontario wholesale
electricity market, the “two schedule” market design and the associated CMSC payment regime.
The IESO administers the wholesale electricity markets in Ontario.39 The IESO operates a
real-time energy market, in which electricity demand and supply are balanced and instructions
are issued to dispatchable generators and loads every five minutes as well as to intertie traders on
an hourly basis. The IESO selects the most economic offers from generators and importers as
well as bids from dispatchable loads and exporters in order to match the supply and consumption
of electricity for each five-minute interval. The outputs of this process include dispatch
quantities and the Market Clearing Price (“MCP”). The simple average of the 12 interval MCPs
in an hour is the Hourly Ontario Energy Price (“HOEP”).40
6.1
Dispatchable Loads
Most users of electricity (also known as loads), even those that are directly connected to the
IESO-controlled grid, are not actively involved in the wholesale market (i.e. they do not submit
bids to the IESO to buy electricity) and the IESO does not (except in emergency conditions)
direct or control the amount of electricity they consume. These customers are referred to as
non-dispatchable loads.
The Market Rules allow loads to become dispatchable and to submit bids in the wholesale
market which indicate the quantity of electricity they wish to consume at particular price levels.
To qualify as a dispatchable load, a facility must be capable of receiving and responding to
dispatch instructions sent every five minutes by the IESO.41 The IESO directs (dispatches) a
39
See, e.g., IESO, Introduction to Ontario’s Physical Markets: An IESO Marketplace Training Publication, online:
http://www.ieso.ca/imoweb/pubs/training/IntroOntarioPhysicalMarkets.pdf.
40
See, e.g., IESO, Overview of the IESO-Administered Markets: An IESO Training Publication, online:
http://www.ieso.ca/imoweb/pubs/training/MarketsOverview.pdf.
41
See, e.g., IESO, Quick Takes QT17: Dispatchable Loads, online:
http://www.ieso.ca/Documents/training/QT17_DispLoads.pdf. A dispatchable load is also eligible to provide
operating reserve to the IESO’s operating reserve market.
PUBLIC VERSION
Products Inc.
30
dispatchable load’s energy consumption based on its bids, market supply and demand and
conditions in the load’s local area.
A dispatchable load may submit only one energy bid for each registered facility for any dispatch
hour, but such a bid may contain “laminations” of up to 20 price-quantity pairs (“P/Q Pairs”).42
The price in each price-quantity pair cannot be greater than the Maximum Market Clearing Price
(“MMCP”) of $2,000/MWh, or lower than the negative MMCP (-$2,000/MWh).
If a load participating in the wholesale market bids at $1,999/MWh, it is treated as dispatchable
by the IESO and will be scheduled at its desired consumption level (unless there is a need to
dispatch it down/off in order to balance demand with supply after all other lower priced
resources have been dispatched). However, if the load bids all or part of its consumption at the
MMCP ($2,000/MWh) in any hour, that quantity is deemed by the IESO to be non-dispatchable
in that hour (and therefore ineligible to be dispatched down).43
When market prices are negative, loads and exporters are paid to consume energy (instead of
paying for their energy) and generators and importers are charged (instead of being paid) to
supply energy. A bid at a negative price means the load is willing to consume energy only if it is
paid to do so. For example, a bid of 20 MW at -$100/MWh indicates the load is willing to
consume 20 MW only if the market price is less than (i.e., more negative) or equal
to -$100/MWh.
Dispatchable loads also submit ramp rates (in MW/minute) which indicate how quickly the load
can change the amount of energy it is consuming. The IESO uses this information to determine
42
IESO, Dispatchable Load Operating Guide, online:
http://www.ieso.ca/imoweb/pubs/training/DispLoadGuide.pdf.
43
IESO, Market Manual 4: Market Operations, Part 4.2: Submission of Dispatch Data in the Real-Time Energy and
Operating Reserve Markets, p. 10. For example, a bid may contain a first lamination of 20 MW for the MMCP of
$2,000/MWh, indicating the load is non-dispatchable for 20 MW and wants to consume regardless of market price.
The bid may also contain a second lamination of 50 MW at $500/MWh, indicating the load is willing to consume a
total of 50 MW (i.e. an incremental 30 MW) so long as the market price is less than or equal to $500/MWh. At any
price above $500/MWh, the load only wants to consume the first lamination of 20 MW.
PUBLIC VERSION
Products Inc.
31
dispatch instructions that a facility can physically follow. A dispatchable load can enter up to
five ramp rates for the laminations comprising its energy bid for each hour.
6.2
The Two-Schedule Market Design
The real-time wholesale energy market is a uniform-price market in which suppliers (generators
and importers) generally receive, and wholesale customers (including dispatchable and
non-dispatchable loads as well as exporters) generally pay, a system-wide market price44 for
electricity irrespective of their location in Ontario.45 The decision to adopt a uniform-price
market, rather than a market in which prices vary by location, has resulted in a “two-schedule”
system in order to deal with differences between the province-wide “market” (or
“unconstrained”) demand/supply and the physical capabilities of the system which results in the
need for the IESO to “constrain” market participants in order to deal with localized
demand/supply imbalances.
Under the two-schedule system, the IESO’s dispatch algorithm is run in two modes for every
five-minute interval of market operation:
·
The “unconstrained mode” ignores most physical limitations of the transmission system
inside Ontario. The outputs are settlement prices and “market schedules” (also referred to
as “unconstrained schedules”) that show the amount of energy that dispatchable facilities
would have been prepared to inject or withdraw if there were no constraints on the
system.
·
The “constrained mode” considers all physical limitations of the grid including
transmission constraints and transmission line losses (“Grid Conditions”). The outputs
are the dispatch instructions that are issued by the IESO and “dispatch schedules” (also
44
The price for generators and dispatchable loads is the MCP for each interval. Non-dispatchable loads pay the
HOEP. Import and export transactions are also paid based on the HOEP, subject to adjustments related to localized
intertie congestion.
45
The description of the two-schedule system in this section and the following subsections is a simplified summary.
For more detail, see IESO, Introduction to Ontario’s Physical Markets, online:
http://www.ieso.ca/imoweb/pubs/training/IntroOntarioPhysicalMarkets.pdf.
PUBLIC VERSION
Products Inc.
32
referred to as “constrained schedules”) that show energy injections and withdrawals for
dispatchable facilities that can actually happen within the physical constraints of the
system.
A dispatchable resource is “constrained on” when the constrained schedule dispatches it to
produce or consume more electricity than is indicated in the market schedule. Conversely, a
dispatchable resource is “constrained off” when the constrained schedule dispatches it to produce
or consume less electricity than is indicated in the market schedule.
The main differences between unconstrained and constrained modes of the dispatch algorithm
are summarized in Table 6-1 and are discussed in more detail in Appendix D (unconstrained
schedule) and Appendix E (constrained schedule).
PUBLIC VERSION
Products Inc.
33
Table 6-1: Differences between the Unconstrained and Constrained Modes
of the IESO’s Dispatch Algorithm
Attribute
Constrained Mode
Unconstrained Mode
Inputs –
Prices and
Quantities
Hourly bids and offers from dispatchable facilities,
self-scheduling generator quantities, intermittent
generation forecasts and a forecast of demand by
non-dispatchable facilities.
Same as constrained mode.
Inputs –
Transmission
Constraints
Includes all transmission constraints and limitations
within Ontario.
Ignores most transmission constraints
within Ontario.
Inputs –
Ramp Rates
Uses ramp rates submitted by market participants.
Uses ramp rates that are three times
faster than the participant-submitted
rates used in the constrained schedule
(the “3x ramp rate multiplier”).
Outputs –
Quantities
A constrained (or dispatch) schedule for each
dispatchable participant that shows energy injections
and withdrawals for a five-minute interval that can
actually happen within the physical constraints of the
transmission system and participant equipment.
The constrained schedule is the basis for the IESO’s
dispatch instructions to dispatchable facilities.
An unconstrained (or market) schedule
for each dispatchable participant that
shows the amount of energy the
participant would be
injecting/consuming in a five-minute
interval given (a) its offer/bid, (b) ramp
rates determined using the 3x ramp rate
multiplier, and (c) the absence of
transmission constraints on a provincewide basis.
Outputs –
Prices
“Nodal Prices” for each injection or withdrawal
node on the Ontario transmission system. These
reflect the marginal cost of supplying an additional
MW at that particular location on the grid, based on
the offers/bids of market participants and constraints
arising from Grid Conditions.
These prices are compared to the bid/offer prices
submitted by participants to determine whether a
dispatchable facility should be constrained on or
constrained off.
Market Clearing Prices for each
five-minute interval.
Supply and
Demand in
Future Intervals
The constrained schedule for a five-minute interval is
determined by considering offers, bids and Grid
Conditions for the current interval and for the next
several intervals. Multi-interval optimization
provides a more efficient dispatch in the current and
future intervals.
The starting point for each new interval is the actual
production or consumption in the most recently
completed interval.
PUBLIC VERSION
These MCPs (or the hourly average of
them - HOEP) are used in the IESO’s
billing and settlement system.
The unconstrained mode considers only
offers and bids for a single five-minute
interval. It does not “look ahead” to
offers and bids for upcoming intervals.
The starting point for each new interval
is the unconstrained schedule for the
most recently completed interval.
Products Inc.
34
6.3
Congestion Management Settlement Credits
CMSC payments are intended to compensate a dispatchable market participant when, based on
the constrained schedule, the IESO instructs it to supply (dispatchable generator or importer) or
consume (dispatchable load or exporter) electricity at an amount that is less profitable for the
participant relative to the operating profit that would have been expected from generating or
consuming at the level indicated for the participant in the market schedule.
6.3.1 The Origin of CMSC Payments
CMSC payments arose from the decision to adopt a uniform-price market and the two-schedule
system. The Market Design Committee, charged with designing Ontario’s electricity market,
proposed such payments to compensate dispatchable facilities for reductions in their operating
profits that resulted from responding to system operator instructions to alter their output or
consumption in order to relieve transmission constraints:
A uniform “market” price (the price is actually administratively
determined) implies a set of corresponding market quantities that
each participant would sell or buy at that uniform market price.
However, transmission constraints may prevent participants from
injecting or withdrawing those corresponding market quantities. In
order to relieve the actual constraints and remain within system
security limits during dispatch, the IMO [now IESO] may have to
direct generators (and dispatchable loads) to produce (consume)
more or less energy than they are willing to produce (consume) at
the uniform price, given the prices each participant has indicated in
its bid or offer. To induce generators and loads to change their
outputs or takes to the required levels, a uniform pricing approach
thus requires the IMO to compensate participants for any
differences between the uniform price and their bids/offers
whenever they are “constrained on” or “constrained off” in order to
relieve transmission constraints.46 (emphasis added)
46
Market Design Committee, Final Report of the Market Design Committee: To the Honourable Jim Wilson,
Minister of Energy, Science and Technology, January 29, 1999, Volume 1, ch. 3, p. 8, online:
http://www.ieso.ca/Documents/mdc/Reports/FinalReport/Volume-1.pdf.
PUBLIC VERSION
Products Inc.
35
6.3.2 Determination of CMSC Payments
The Market Rules established CMSC payments as compensation for reduced operating profits
that result from responding to dispatch instructions to produce or consume at a level different
than the market schedule:
Dispatch instructions provided by the IESO to market participant
'k' will sometimes instruct k to deviate from its market schedule in
ways that, based on market participant k's offers and bids, imply a
change to market participant k's net operating profits relative to the
operating profits implied by market participant k's market
schedule. When this occurs and market participant k responds to
the IESO's dispatch instructions, market participant k shall, subject
to Appendix 7.6 of Chapter 7, receive as compensation a
settlement credit equal to the change in implied operating profits
resulting from such response, calculated in accordance with
section 3.5.2.47 (emphasis added)
6.3.2.1 Formula for Calculation of CMSC
As indicated in Table 6-2, the CMSC payment for a dispatchable load in any five-minute interval
is effectively calculated as the difference between its bid price and the MCP, multiplied by the
difference between its unconstrained schedule and constrained schedule (or in certain
circumstances the load’s actual consumption) quantities.48
Table 6-2: Simplified CMSC Formulas
Simplified Constrained-Off CMSC Formula:
Payment = [Bid Price – MCP] x [Unconstrained MW– Constrained MW]
Simplified Constrained-On CMSC Formula:
Payment = [MCP – Bid Price] x [Constrained MW – Unconstrained MW]
47
48
Market Rules, Chapter. 9, Section 3.5.1.
A detailed explanation of the actual formulas is provided in Appendix F.
PUBLIC VERSION
Products Inc.
36
For constrained-off CMSC, as a dispatchable load’s bid price increases, all else being equal, so
too does the CMSC payment. Conversely, for constrained-on CMSC, as a dispatchable load’s bid
price decreases the CMSC increases. In both cases, as the difference between the quantities to be
consumed in the constrained and unconstrained schedules grows so too does the CMSC payment.
CMSC payments (or, in rare cases, negative amounts which must be repaid by the market
participant) are determined for each five-minute interval (subject to certain “clawback” rules
which are discussed below).
The IESO dispatches generators and loads based on the Nodal Price at each injection or
withdrawal node on the Ontario transmission system. These Nodal Prices may differ from the
MCP, sometimes dramatically. Depending on the relationship between the Nodal Price and a
load’s bid price, the IESO can instruct a load to consume more or less energy than the amount
that appears in the load’s market schedule. Even if a load’s bid price is considered to be
economic in the market schedule (i.e., higher than the MCP), it will be required to curtail
consumption when the Nodal Price at its withdrawal node is higher than its bid price. Similarly,
it will be required to consume energy when its bid price exceeds the Nodal Price at its
withdrawal node, even if it is uneconomic in the market schedule (i.e., its bid price is lower than
the MCP).
The CMSC payments that will be made to a load that is constrained off or constrained on by the
IESO are illustrated in Table 6-3.
PUBLIC VERSION
Products Inc.
37
Table 6-3: Illustration of CMSC Payments Arising from Differences in the Unconstrained and
Constrained Modes of the IESO’s Dispatch Algorithm
Load Constrained Off
Load Constrained On
10 MW
2 MW
Load's bid price
$100/MWh
-$900/MWh
Market Clearing Price (MCP)
$30/MWh
$30/MWh
Nodal Price at load's
withdrawal node
$150/MWh
-$1000/MWh
How many MWs would be
scheduled based on the
uniform, province-wide MCP?
10 MW
(because $100 bid > $30 MCP)
0 MW
(because -$900 bid < $30 MCP)
How many MWs will be
included in the constrained
schedule?
0 MW
(because $100 bid < $150 Nodal
Price)
2 MW
(because -$900 > -$1000 Nodal Price)
$700
($100 bid - $30 MCP) x
(10 MW - 0 MW)
$1860
($30 MCP - (-$900 bid)) x
49
(2 MW - 0 MW)
Load's bid quantity
CMSC Payment
6.3.2.2 Relationship Between Marginal Benefit of Consumption and Operating Profits
The Marginal Benefit of Consumption is the incremental net revenue expected to result from
increasing production by consuming an additional MW of electricity. “Net revenue” is the
revenue expected to result from selling the additional output less variable costs of production
(other than electricity). The change in operating profit is the incremental net revenue less the cost
of electricity for the additional MW. A load normally would not be prepared to pay more for
electricity than the Marginal Benefit of Consumption. If it did so, the cost of the extra MW
would exceed the incremental net revenue from increasing output (i.e., its operating profits
would be reduced). The Marginal Benefit of Consumption may also be used to measure the lost
net revenues (again, before considering electricity costs) when a load consumes one less MW of
electricity. A load normally would not reduce its consumption if the Marginal Benefit of
Consumption exceeded the price of electricity.
49
However, because the dispatchable load is constrained on, it must also pay for the power it consumed at the MCP
of $30/MWh. As a result, the net payment to the dispatchable load is $1,800.
PUBLIC VERSION
Products Inc.
38
While there is no Market Rule that requires a bid price submitted by a dispatchable load to
reflect the load’s Marginal Benefit of Consumption, the Market Rules relevant to CMSC assume
that a bid price will reflect a dispatchable load’s actual benefit of consumption:
The dispatch scheduling and pricing process shall be a
mathematical optimisation algorithm that will determine optimal
schedules for each time period referred to in section 2.1.1, given
the bids and offers submitted and applicable constraints on the use
of the IESO-controlled grid. Marginal cost-based prices shall also
be produced and, for such purpose, offer prices shall be assumed to
represent the actual costs of suppliers and bid prices shall be
assumed to represent the actual benefits of consumption by
dispatchable load facilities.50 (emphasis added)
In other words, the CMSC calculation assumes that the bid price submitted by a dispatchable
load would reflect the load’s Marginal Benefit of Consumption. The load’s operating profit is
assumed to be reduced whenever the load is directed by the IESO to consume less “cheap”
power (MCP < load’s bid price) than it otherwise would. Similarly, when market prices are
“expensive” from the load’s perspective (MCP > load’s bid price), the load’s operating profit is
assumed to be reduced whenever the IESO requires the load to consume more than it otherwise
would.
6.3.3 Self-Induced CMSC Payments
The Market Design Committee’s Report and the Market Rules clearly indicate that three
conditions should exist for a CMSC payment to be made:
(i)
the reason for constrained-on or constrained-off dispatch instructions
relates to Grid Conditions (i.e., the IESO instructs a load to consume
electricity in larger or smaller amounts than the economics of the load’s
bid would otherwise dictate in order to relieve transmission constraints
and remain within system security limits);
50
Market Rules, Appendix 7.5, Section 2.3.1.
PUBLIC VERSION
Products Inc.
39
(ii)
the load would have consumed a different amount of energy absent the
constrained-on or constrained-off dispatch instruction, and it earns lower
operating profits by following the IESO’s instruction; and
(iii)
the amount of the CMSC payment should be limited to the amount
necessary to provide compensation for operating profit reductions that are
linked to the two foregoing conditions.
Although these conditions appear to be straightforward and sensible, the Market Rules and the
IESO’s settlement tools have been formulated in a manner which may allow CMSC payments to
arise in other situations and may result in the market participant receiving payments that exceed
compensation for reduced operating profits arising from responses to dispatch instructions
caused by Grid Conditions.
In particular, a dispatchable load may be able to self-induce CMSC payments. “Self-induce”
refers to the ability of the market participant to bring about the outcome, in this case the CMSC
payment, through its own voluntary actions. The market participant can self-induce a CMSC
payment by creating quantity differences between the unconstrained or constrained schedules
either through its offer/bid submissions or its consumption behaviour. Actions that self-induce
CMSC payments may or may not constitute gaming, depending on whether the market
participant exploited a market defect, such as bidding above Marginal Benefit of Consumption
during self-induced ramping hours, to its profit or benefit and to the expense or disadvantage of
the market (see the analytical framework for gaming set out in Section 5.5). Both types of
actions, self-inducing quantity differences and bidding at a price that does not reflect the
Marginal Benefit of Consumption, are examined in relation to Bowater (see Section 7) and
Abitibi (see Section 8).
6.3.4 Clawback of Certain CMSC Payments
CMSC payments are automatically determined by the IESO’s settlement tools for every fiveminute interval in which there are differences between the unconstrained and constrained
schedules. The CMSC formulas do not distinguish between system-induced CMSC that arises
PUBLIC VERSION
Products Inc.
40
from Grid Conditions and self-induced CMSC. However, in 2003 the Market Rules were
amended to allow the IESO to either avoid making or completely recover self-induced
constrained-off (but not constrained-on) CMSC payments from dispatchable loads under the
following circumstances:
A registered market participant for a constrained off facility is not entitled to a
congestion management settlement credit determined in accordance with section
3.5.2 as the result of that registered facility’s own equipment or operational
limitations, if:
3.5.1A.1 a dispatchable load facility does not fully or accurately respond
to its dispatch instructions; or
3.5.1A.2 the ramping capability of a dispatchable load facility, as
represented by the ramp rate set out in the offers or bids, is below the
threshold for the IESO to modify dispatch instructions and thereby
prevents changes to the dispatch;
and then the IESO may withhold or recover such congestion management
settlement credits . . . . 51
The IESO initially relied on manual processes to identify self-induced CMSC payments that
should be recovered. In 2007, the IESO introduced an automated approach to CMSC recovery,
and documented the procedures used to calculate the amount of self-induced CMSC payments
that may be clawed back under the Market Rules. Those procedures contain the four criteria
(referred to as the “Business Rules”) which are applied by the IESO to recover constrained-off
CMSC payments from dispatchable loads (see Appendix H).52
The Business Rules have resulted in the clawback of a significant amount of self-induced
constrained-off CMSC payments from dispatchable loads. During the Relevant Period, gross
CMSC payments to all dispatchable loads in Ontario were $44.1 million, of which $19.1 million
(43%) was clawed back under the Business Rules. If the Business Rules were perfectly
effective, the net CMSC payments of $25 million should represent the amount required to
51
52
Market Rules, Chapter 9, Section 3.5.1A, Issue 22.0.
IESO, Market Manual 5, Part 5.5: Physical Markets Settlement Statements (Issue 38.0), s. 1.6.8.
PUBLIC VERSION
Products Inc.
41
compensate dispatchable loads for following IESO dispatch instructions in response to Grid
Conditions. However, as will be seen in Sections 7 and 8 of this Report, the Business Rules do
not operate so as to recover all constrained-off CMSC payments in excess of compensation for
such operating profit reductions, particularly in situations involving self-induced ramping,
deviations from dispatch instructions or bid prices that do not reflect the load’s Marginal Benefit
of Consumption. Moreover, they do not address constrained-on CMSC payments such as those
arising from the negative-price bidding strategy used by Abitibi in certain hours (see Section
8.5).
PUBLIC VERSION
Products Inc.
42
7.
BOWATER’S CONDUCT IN RESPECT OF THE THUNDER BAY FACILITY
This section contains the Panel’s assessment of whether Bowater engaged in gaming in relation
to CMSC payments made in respect of the Thunder Bay Facility. The introductory sections
(Sections 7.1 and 7.2) describe the constrained-off CMSC payments received and Bowater’s
typical operating pattern for the Facility, including its bidding strategy and ramping pattern. The
subsequent sections (Sections 7.3 to 7.6) consider the four elements of the gaming framework set
out in Section 5.5, namely whether there were market defects which were exploited by Bowater
to its profit or benefit and to the expense or disadvantage of the market. The Panel concludes
(Section 7.7) that four of Bowater’s behaviours constituted gaming.
7.1
CMSC Payments to Bowater
Between February and August 2010, Bowater received approximately $12.3 million in net
CMSC payments (after clawbacks). All of the payments were for being constrained off.
Although much of these CMSC payments were self-induced, as will be described in this section,
the Business Rules only recovered a portion of such payments. Table 7-1 summarizes the CMSC
payments to Bowater during the Relevant Period, including clawbacks.
Table 7-1: Gross and Net CMSC Payments to Bowater for the Thunder Bay Facility
February – August 2010
($000)
Month
February
March
April
May
June
July
August
Total
Gross ConstrainedOff
CMSC
2,202
3,040
3,549
3,463
2,438
Clawback of
Constrained-off
CMSC
797
532
947
1,125
642
1,159
1,796
292
977
867
818
$17,647
$5,312
$12,334
PUBLIC VERSION
Net Constrained-Off
CMSC
1,405
2,508
2,602
2,339
1,796
Products Inc.
43
7.2
Typical Operating Pattern
The Thunder Bay Facility’s energy consumption pattern in 2010 differed from historical
patterns. Pursuant to a 2009 agreement with the Ontario Power Authority (OPA) which
preceded the reopening of the Thunder Bay Facility (the “DR2 Agreement”),53 Bowater agreed
to reduce its energy consumption during “peak hours” when demand for energy was highest.
Under the terms of the DR2 Agreement, Bowater agreed to curtail ● MW during its “On-Peak
Contract Period” of 8:00 a.m. to 6:00 p.m. on weekdays.54 Therefore, each weekday, the load
was ramped down to a low level at or near 8:00 a.m., and ramped back up to a high level starting
at or near 6:00 p.m. The DR2 Agreement did not apply to operations on the weekend.
7.2.1 Bidding Strategy
The Panel’s review of Bowater’s historical bid practices when it was dispatchable prior to
September 2006 revealed that the use of a $●/MWh bid price had been limited to a small portion
of the Thunder Bay Facility’s load. Between November 2003 and July 2005, Bowater regularly
submitted bids of $2,000/MWh (i.e. making the load non-dispatchable) or $●/MWh for up to ●
MW of its total load, and submitted bid prices of between $●/MWh and $●/MWh for the balance
of its load (during this time period the total load typically fluctuated between ● MW and ● MW).
After July 2005 and before becoming non-dispatchable in September 2006, Bowater did not use
bids of $●/MWh; instead it generally submitted bids between $●/MWh and $●/MWh during
normal operations as well as when ramping up and ramping down (except for bids of $●/MWh at
certain times).
Upon returning to the wholesale market as a dispatchable load, from February 2010 until June
2010, Bowater consistently submitted [an extremely high bid price]. Bowater submitted this bid
price for all hours, including those hours when it signalled its intention to ramp by changing its
bid quantity from one hour to the next. From July 2010 through August 2010, after discussions
53
DR2 Contract between Bowater Canadian Forest Products Inc. and Ontario Power Authority, October 21, 2009.
54
Ibid.
PUBLIC VERSION
Products Inc.
44
with the MAU, Bowater generally maintained a bid price of $●/MWh during daytime operations,
but used a bid price of $●/MWh during evening hours and when making changes to its quantity
bids that signalled its intention to ramp up or down. The same bidding strategy was used on both
weekdays and weekends.
7.2.2 Ramping Pattern
During weekdays in the Relevant Period, the Thunder Bay Facility normally ramped down to
approximately ● MW by 8:00 a.m., and then ramped up to approximately ● MW usually starting
at 6:00 p.m. The total morning and evening consumption change exceeded the ● MW
commitment in the DR2 Agreement.
On weekends during the Relevant Period, the Thunder Bay Facility typically performed one
ramp down and ramp up. Bowater indicated that these ramps were for the purpose of inventory
management.55
To implement a ramp down (normally in Hour Ending (“HE”) 7 on weekdays), Bowater
decreased its quantity bid from the night-time level (● MW until May 12, 2010 and ● MW
thereafter) to the day-time level of ● MW.
To implement a ramp up (normally in HE 19 on weekdays), Bowater increased its quantity bid
from the day-time level of ● MW: first through a small step up to ● MW, followed by a change
in the next hour from ● MW to the night-time level (● MW or ● MW).
During the Relevant Period, Bowater used ramp rates which ramped the Thunder Bay Facility up
or down in three stages, as summarized in Table 7-2. The relationship between Bowater’s ramp
rates and the CMSC payments it received is discussed in Sections 7.4.4 and 7.4.5.
55
Responses to RFI, B.13, p. 2.
PUBLIC VERSION
Products Inc.
45
Table 7-2: Ramp Rates for the Thunder Bay Facility
(MW and MW/min)
RAMP DOWN
Machine
First and Second
Mainline Refiners
Rejects Refiner and
Mainline Refiner Motors
Auxiliaries
RAMP UP
MW Range
Ramp Rate
● (or ●) to
●
● to ●
●–●
MW/min
● to ●
●MW/min
●–
●MW/min
Machine
Auxiliaries
MW Range
Ramp Rate
● to ●
● MW/min
Rejects Refiner and
Mainline Refiner
Motors
First and Second
Mainline Refiners
● to ●
●–●
MW/min
● to ● (or ●)
●–●
MW/min
Bowater utilized different ramp up sequences in 2010 than during its previous history as a
dispatchable load.56 In 2006, the rejects refiner was started 10-15 minutes prior to the hour, the
first pair of mainline refiners was started on the hour, and the second pair of mainline refiners
was loaded the following hour. The usual ramp-up profile in 2006,57 in combination with the
submitted ramp rates and bid prices, resulted in minimal CMSC payments (on the order of a few
hundred dollars per ramp). In 2010, Bowater ramped its two mainline refiners and the rejects
refiner in the same hour58 with a typical ramp-up profile that triggered upwards of $20,000 in net
CMSC payments per ramp up.
7.2.3 Constrained-off CMSC Payments During Ramping
As discussed in Section 6.3.4 and Appendix G, a dispatchable load can initiate ramping through
a “self-induced dispatch” by changing the prices and/or quantities that it bids (i.e., its P/Q Pairs).
Both the market and the constrained schedules will change to allow the load to ramp to its
desired new level of consumption. The changes to the load’s dispatch are caused by the ramping
56
During ramp down Bowater utilized similar sequences in 2010 as it did in 2006. On ramp down, one pair of
mainline refiners was shutdown followed 5 minutes later by the second mainline refiner and then the rejects refiner
10 minutes later. The auxiliary loads were shut down over the following 10-15 minute period, and the recycle plant
was then started-up as quickly as possible. Responses to RFI, B.2, p. 4.
57
Bid data in 2006 shows ramp ups were generally scheduled from ● MW to ● MW in one hour, and from ● MW to
● MW in the next hour.
58
PUBLIC VERSION
Products Inc.
46
decision manifested in the bid of the dispatchable load, not by Grid Conditions. Nevertheless,
CMSC payments will be made if the schedule quantities diverge as a result of a change in the
price or quantity bid by the load. The resulting CMSC payments are self-induced.
During the Relevant Period, Bowater induced the desired changes in its energy consumption by
changing its quantity bid between one hour (“h”) and the following hour (“h+1”). Because the
Thunder Bay Facility requires more than one five-minute interval to ramp to its new
consumption level, the dispatches in the constrained schedule diverged from the market
schedule.59
An illustration of ramping behaviour at the Thunder Bay Facility can be seen in the consumption
pattern of Friday, May 14, 2010, when the facility earned over $73,000 in CMSC (net of
clawbacks). Bowater bid $●/MWh in every hour and submitted the ramp rates noted previously
in Table 7-2. The ramping of consumption, and the CMSC payments made during the ramping
periods, are shown in Figure 7-1.
59
During a ramp down at the Thunder Bay Facility, the constrained schedule will dispatch the Facility down (i.e.
ramp) to its new consumption level by the end of hour h, while the market schedule will ramp the Facility down in
the first interval of hour h+1. This is because the dispatch algorithm will not dispatch a facility in the constrained
schedule above its maximum submitted quantity in hour h+1 or in a way that is inconsistent with the actual ramp
rates of the facility, which is not the case in the unconstrained schedule. During a ramp up, the constrained schedule
will normally dispatch a facility to begin ramping towards its new consumption level in the first interval of hour
h+1, while the market schedule will begin ramping at the same time but using the 3x times ramp rate multiplier. (See
Section 6.3 as well as Appendix D, Appendix E and Appendix G). In either case, the result is a discrepancy between
the market and constrained schedules during the ramping period. This in turn triggers constrained-off CMSC
payments because the quantity in the constrained schedule falls below the market schedule quantity during the ramp
intervals.
PUBLIC VERSION
Products Inc.
47
Figure 7-1: Sample Ramping Pattern and CMSC Payments for the Thunder Bay Facility
May 14, 2010
(MW and $)
7.2.3.1 CMSC on Ramp Down
To accommodate Bowater’s morning change in bid quantity on May 14, 2010, the dispatch
algorithm began to dispatch the Thunder Bay Facility down beginning in interval 8 of HE 6
using the submitted ramp-down rates. Accordingly, it was dispatched from ● MW to ● MW over
five intervals. Throughout this period the Facility’s market schedule remained at ● MW. To
calculate CMSC, the IESO settlement tool took the quantity difference between (i) the market
schedule and (ii) the greater of the constrained schedule and actual consumption quantities
during each interval of the ramp period, and multiplied each quantity difference by (iii) the bid
price less (iv) the MCP for the interval (which was in the vicinity of $35/MWh). In the result,
over $50,000 in CMSC was paid in respect of this particular ramp down event.
Table 7-3 shows the total energy charges and CMSC payments to Bowater for each interval in
HE 6 on May 14, 2010 when the Thunder Bay Facility was ramping down. The net CMSC
PUBLIC VERSION
Products Inc.
48
payments ($52,908) were substantially larger than the energy charges ($5,155), such that
Bowater was receiving nearly $48,000 while consuming the amount of energy it wanted to
during its self-induced ramp down.
Table 7-3: Energy Charges and CMSC Payments on a Typical Ramp Down
of the Thunder Bay Facility
May 14, 2010, HE 6
(MW, $/MWh and $)
Interval
Unconstrained
Schedule
(MW)
Constrained
Schedule
(MW)
Actual
Consumption
(MW)
Energy
Charges*
($)
MCP
($/MWh)
35.32
33.38
33.61
33.38
34.55
35.17
35.32
35.32
35.33
35.33
35.37
35.82
●
●
●
●
1
●
●
●
●
2
●
●
●
●
3
●
●
●
●
4
●
●
●
●
5
●
●
●
●
6
●
●
●
●
7
●
●
●
●
8
●
●
●
●
9
●
●
●
●
10
●
●
●
●
11
●
●
●
●
12
Total
$5,155
** None of the Business Rules applied to claw back CMSC payments on the ramp down.
Bid
Price
($/MWh)
Net
CMSC**
($)
●
●
●
●
●
●
●
●
●
●
●
●
4,376
9,539
11,538
13,729
13,726
$52,908
7.2.3.2 CMSC on Ramp Up
To accommodate Bowater’s change in bid quantity for the evening of May 14, 2010, the dispatch
algorithm dispatched the Thunder Bay Facility up beginning in interval 1 of HE 19 using the
submitted ramp up rates. Accordingly, it was dispatched from ● MW to ● MW over five
intervals. 60 The 3x ramp rate multiplier resulted in the market schedule moving to ● MW in
interval one, and then to ● MW for the remaining intervals. To calculate CMSC, the IESO
settlement tool took the quantity difference between (i) the market schedule and (ii) the greater
60
The Facility was dispatched up from ● MW to ● MW in the next hour, HE 20. The staged ramp up corresponded
to the staged increase in Bowater’s submitted bid quantities.
PUBLIC VERSION
Products Inc.
49
of the constrained schedule and actual consumption quantities for each interval during the ramp
period, and multiplied each quantity difference by (iii) the $●/MWh bid price less (iv) the MCP
for the interval (which was approximately $33-35/MWh). In the result, over $20,000 in net
CMSC payments were made in relation to this particular ramp up event.
Table 7-4 shows the total energy charges and CMSC payments to Bowater for each interval in
HE 19 on May 14, 2010 when the Thunder Bay Facility was ramping up. The net CMSC
payments ($20,075) were substantially larger than the energy charges ($5,132), such that
Bowater was receiving nearly $15,000 while consuming the amount of energy it wanted to
during its self-induced ramp up.
Table 7-4: Energy Charges and CMSC Payments on a Typical Ramp Up
of the Thunder Bay Facility
May 14, 2010, HE 19
(MW, $/MWh and $)
Interval
Unconstrained
Schedule
(MW)
Constrained
Schedule
(MW)
Actual
Consumption
(MW)
Energy
Charges*
($)
MCP
($/MWh)
35.33
35.33
35.35
35.34
35.38
35.37
35.34
35.36
35.35
35.33
35.34
35.34
Bid
Price
($/MWh)
Net
CMSC **
($)
3,302
9,818
6,955
●
●
●
●
●
1
●
●
●
●
●
2
●
●
●
●
●
3
●
●
●
●
●
4
●
●
●
●
●
5
●
●
●
●
●
6
●
●
●
●
●
7
●
●
●
●
●
8
●
●
●
●
●
9
●
●
●
●
●
10
●
●
●
●
●
11
●
●
●
●
●
12
Total
$5,132
$20,075
*Includes Global Adjustment and Uplift charges.
**CMSC payments of $4,402 in interval 4 and $1,477 in interval 5 were clawed back under Business Rule 3.
7.2.3.3 Representative Pattern of Operation
The bids, ramping and consumption on May 14, 2010 are typical of the Thunder Bay Facility’s
operating pattern on weekdays during the Relevant Period. On weekends, the Facility was not
PUBLIC VERSION
Products Inc.
50
subject to the requirements of the DR2 Agreement. The Facility typically ramped down once on
the weekend for a brief period of time (approximately 3 hours) before ramping back up.
Although weekend ramp downs and ramp ups occurred at different times of day, they typically
triggered CMSC payments in a similar manner as illustrated by the May 14th example.
The amount of the CMSC payments on any particular weekday or weekend ramp varied
primarily in response to variations in the magnitude of the ramp, actual consumption relative to
the constrained schedule, and the MCPs during the applicable ramp intervals. A summary of
Bowater’s five largest CMSC payment days at the Thunder Bay Facility during the Relevant
Period is provided in Appendix I. Graphs comparable to Figure 7-1 which show the hourly
constrained and unconstrained schedules, actual consumption and CMSC payments are also
included in Appendix I.
7.3
Defects in Market Rules or Procedures
Even before the Ontario electricity market opened in 2002, the Market Design Committee and
the MSP were both concerned that CMSC payments that more than compensate a market
participant for any reduction in its operating profits could be self-induced and would be contrary
to the overall purpose of the CMSC framework. Moreover, the MSP expressed concern that the
CMSC regime was conducive to gaming 61 and the Market Design Committee suggested that
“rules be developed to discourage gaming of side payments.”62
One defect that can be subject to exploitation, and was of particular concern to the MSP, relates
to the formula used for calculating CMSC payments. The formula determines the implied change
in the operating profit for a dispatchable load based on the expectation that its bid reflects its
Marginal Benefit of Consumption. As previously noted, the Market Rules contain the
61
Independent Electricity Market Operator, The Market Surveillance Panel In Ontario’s Electricity Market:
Monitoring, Investigating and Reporting – Backgrounder, April 2002, online:
http://www.ontla.on.ca/library/repository/mon/4000/10306902.pdf, p. 13.
62
Market Design Committee, Second Interim Report of the Market Design Committee: To the Honourable
Jim Wilson, Minister of Energy, Science and Technology, June 30, 1998, online:
http://www.ieso.ca/imoweb/historical_devel/MDC/Reports/InterimReport2/2ndRept.pdf., p. 9 of the Appendix and
p. 3-15.
PUBLIC VERSION
Products Inc.
51
assumption but not a requirement that bids will reflect the Marginal Benefit of Consumption. If a
bid does not reflect the Marginal Benefit of Consumption, but instead is higher, then any
constrained-off CMSC payment would exceed what is required to compensate a dispatchable
load for operating profit reductions as a result of following dispatch instructions that are caused
by Grid Conditions.
The interface between the CMSC regime and the two-schedule system also has defects that can
be exploited in relation to quantity differences. As noted in Section 6.3 as well as Appendix D
and Appendix E, there are differences between the optimization and ramping processes in the
unconstrained and constrained modes of the dispatch algorithm. A market participant may be
able to use ramping decisions or deviations from its constrained dispatch schedule to self-induce
quantity differences that give rise to CMSC payments. Such payments are not compensating for
reductions in operating profits caused by responding to Grid Conditions, but are an unintended
consequence of the CMSC regime.
Another defect that can be subject to exploitation relates to the IESO’s procedures for recovering
(or “clawing back”) CMSC payments that go beyond compensation for changes in operating
profits arising from responding to constrained instructions that were based on Grid Conditions.
As noted in Section 6.3.4 and Appendix H, the Business Rules and processes developed by the
IESO to claw back payments in various circumstances do not recover all of the CMSC payments
that are self-induced, either because they do not apply in certain situations, they fail to identify
all instances of inappropriate CMSC payments, or they allow payments which exceed the amount
needed to compensate for operating profit reductions. Bowater’s self-induced CMSC payments
are an example of payments that were not covered by the Business Rules.
PUBLIC VERSION
Products Inc.
52
Finding #1 (Market Defects Related to Constrained-off
CMSC):
The CMSC rules, formulas and clawback procedures that existed during the
Relevant Period allowed a dispatchable load to receive constrained-off
CMSC payments that exceeded the amount required to compensate for
reductions in operating profits arising from responses to dispatch
instructions caused by Grid Conditions.
7.4
Exploitation of Constrained-Off CMSC
As set out in Section 5.5, an essential element of gaming is that the market participant engages in
activity which exploits a market defect. The Panel considers that exploitation may exist where
the market participant had some level of intention, knowledge or awareness of an opportunity
arising from the market defect.63 In order to determine whether Bowater exploited defects in the
CMSC regime, the Panel examined the development of its ramping strategy and the following
specific behaviours, each of which contributed to the large constrained-off CMSC payments
received during ramp periods:
(i)
Bowater submitted an extremely high bid price for ramping hours, which
increased the amount of the CMSC payment for any difference between
the unconstrained and constrained schedule quantities (see Section 7.4.2).
(ii)
Between February and May 2010, Bowater raised its bid quantity above
the level that the Thunder Bay Facility was generally consuming, which
increased the quantity differences used to calculate the CMSC payments
(see Section 7.4.3).
63
See Market Surveillance Panel, Report on an Investigation into Possible Gaming Behaviour Related to Infeasible
Import Transactions by TransAlta Energy Marketing Corp. on the Manitoba-Ontario Intertie, Investigation
No. 2011-02, October 22, 2012, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MSP_Report_Investigation_TranAlta_20121022.pdf,
p. 19.
PUBLIC VERSION
Products Inc.
53
(iii)
Bowater bid such that its entire ramp down would occur within a single
hour, which had the effect of prolonging the period over which a quantity
spread existed between the unconstrained and constrained schedules (see
Section 7.4.4).
(iv)
On ramp downs, Bowater often ramped faster than submitted ramp rates,
which reduced its actual consumption and increased the quantity
differences used to calculate CMSC payments (see Section 7.4.5).
(v)
Bowater occasionally failed to ramp up or down in accordance with its bid
and dispatch instructions, which increased the quantity differences that
give rise to CMSC payments (see Section 7.4.6).
(vi)
Bowater periodically deviated significantly from its dispatch instructions
in non-ramp hours, which triggered CMSC payments that were not always
clawed back under IESO Business Rules (see Section 7.4.7).
7.4.1 Development of the Ramping CMSC Strategy
Prior to re-entering the wholesale market as a dispatchable load in February 2010, Bowater
developed a detailed strategy to maximize CMSC payments during the ramping of the Thunder
Bay Facility. Personnel at the Thunder Bay Facility worked closely with personnel at Abitibi’s
Fort Frances Facility who had experience with the relationship between ramping and CMSC
payments. The key personnel involved in the development of the ramping strategy included:
PUBLIC VERSION
Products Inc.
54
— [Senior Bowater Personnel #3]: [Senior Bowater Personnel #3’s position]
— [Senior Bowater Personnel #5]: [Senior Bowater Personnel #5’s position]
— [Senior Abitibi Personnel #2]: [Senior Abitibi Personnel #2’s position]64
In particular, [Senior Abitibi Personnel #2] worked with [Senior Bowater Personnel #5] to
develop detailed strategies, including bid laminations and ramping sequences, that would
increase the Thunder Bay Facility’s CMSC payments. The following email exchange is an
example:
From:
[Senior Bowater Personnel #5]
To:
[Senior Abitibi Personnel #2]
Dated:
September 11, 2009 09:23 AM
Subject:
Ramp[s] rates and CMSC Payments
[Senior Abitibi Personnel #2]:
Our previous bidding strategy for the TMP [thermo-mechanical
pulpmill] load used a down ramp rate of ●MW/min. The bulk of
the ● MW to TMP load that is shutdown is the refiner loads. There
are two lines of TMP each of about ● MW and a rejects refiner that
is about ●MW. The balance of ●-● MW is for plant auxiliaries that
take 20 to 30 minutes after the refiners are shut down. See the
attached power point file for the shutdown timing that we were
using during the DR2 transition program to ensure we were in
compliance. From this info can you see any potential for CMSC
payments and what would we have to do differently to maximize
these payments.65 (emphasis added)
64
The main Bowater, Abitibi and affiliated company personnel involved in the activities discussed in this Report are
listed in Appendix C.
65
PUBLIC VERSION
Products Inc.
55
[Senior Abitibi Personnel #2] replied as follows:
From:
To:
Dated:
September 11, 2009 12:51 PM
Subject: Ramp Rates and CMSC
The attached is quick look at the potential for CMSC when moving
the load. Notes: You can have different ramp rates applied to your
●MW load. A while ago the IESO moved us from 12x ramp rate to
3x. We should have a discussion on the impacts of this, however it
should not be documented in an email.66 (emphasis added)
The document attached to the above email contained ramp rates, calculations of the dispatch and
market schedules, and the resulting CMSC payments arising from ramping. Abitibi also sent
sample ramps and data “showing actual bids entered with ramp rates, the resulting dispatch
instructions with time stamp and the load prior to dispatch… for a start-up and shut down”67
from the Fort Frances Facility, to aid the Thunder Bay Facility in “forecasting the constrained
schedule correctly.”68 Before and in the weeks after the Thunder Bay Facility became
dispatchable, [Senior Abitibi Personnel #2] would review “ramp rates and start-up/shut down
strategies with respect to CMSC”69 with Thunder Bay Facility personnel, and provide input with
the objective of maximizing CMSC. For example, [Senior Abitibi Personnel #2] advised that
increasing the quantity bid from ● MW to ● MW would increase CMSC payments by
approximately $4,000 per start up.70 [Senior Abitibi Personnel #2] also advised that it was
desirable to have the actual metered consumption quantity below the dispatch quantity when
ramping, and in one email informed Thunder Bay Facility personnel that if they “stayed under
66
Email from [Senior Bowater Personnel #5], January 7, 2010. Responses to RFI, B.2.12.
68
Email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], January 12, 2010. Responses to RFI,
B.2.12.
69
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], January 21, 2010. Responses to
RFI, B.2.24.
70
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #5], February 12, 2010. Responses to
RFI, B.2.15. This issue is discussed more fully in Section 7.4.3 below.
67
PUBLIC VERSION
Products Inc.
56
[their] dispatch this morning’s ramp would pay out $57K.”71 In learning of the large CMSC
amount on the first day of operation of the Thunder Bay Facility as a dispatchable load on
February 8, 2010, [Senior Abitibi Personnel #2] advised to “wait for comment from the IESO
and stay quiet. We are following market rules”.72
Bowater’s ramping CMSC strategy was known to senior management. During the period of
development before the Thunder Bay Facility became dispatchable on February 8, 2010, ABI
personnel were advised of and assisted with the strategy. For example, [Senior AbitibiBowater
Inc Personnel #2], reviewed and commented on a proposed analysis for shutting down and
starting the thermo-mechanical pulpmill or TMP at the Thunder Bay Facility which calculated
“CMSC (based on actual load)” and “Optimized CMSC payment (actual load ≤ constrained
schedule).”73 [Senior AbitibiBowater Inc Personnel #2] commented as follows: “I have not
looked at this for a while… I thought in ramping up you were allowed to keep 3 intervals in total
(the first three), i.e. your total CMSC credit in ramping up would be limited to $406.25. Or
maybe because your ramping up is so spread out, the tools will not claw back the CMSC in the
second hour”.74 [Senior AbitibiBowater Inc Personnel #2]’s comments acknowledged the
development of the Thunder Bay Facility’s CMSC ramping strategy, and indicate that ABI was
aware of Bowater’s strategy.
In Fall 2009, prior to the Thunder Bay Facility becoming a dispatchable load, personnel at the
Thunder Bay Facility prepared a PowerPoint presentation for two Vice-Presidents of ABI
([AbitibiBowater Inc Executive #4] and [AbitibiBowater Inc Executive #3]) entitled “Thunder
Bay 2010 Power Cost – October 1st, 2009”.75 The presentation discussed the forecast impact of
71
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #5], September 23, 2009, Responses to
RFI, B.2.9, and email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #5], February 8, 2010,
72
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #3], February 23, 2010. Responses to
RFI, B.16.35.
73
This analysis is reproduced in Appendix L.
74
Email from [Senior AbitibiBowater Inc Personnel #2] to [Senior Bowater Personnel #5], January 26, 2010.
Responses to RFI B.2.26.
75
Attachment to email from [Senior Bowater Personnel #2] to [Senior Bowater Personnel #5] and [Senior Bowater
Personnel #3], September 28, 2009. Responses to RFI, B.3.6. Relevant excerpts are reproduced in Appendix J.
PUBLIC VERSION
Products Inc.
57
several items, including CMSC payments, on Bowater’s 2010 power cost. The first graph in the
presentation (reproduced as Figure 7-2 below) contains a projection that Bowater would receive
CMSC payments equivalent to $10.00/MWh on its 2010 energy consumption (which was revised
upward from a prior estimate of $0.95/MWh76).
Figure 7-2: Excerpt from “Thunder Bay 2010 Power Cost – October 1st, 2009”,
a presentation by Bowater to AbitibiBowater Inc.
Figure Redacted – Contains Confidential Information
Another slide in the presentation correctly noted that “[t]he market rules assume that participants
place bids and offers based on their marginal cost and benefit.”77 Despite Bowater’s knowledge
of what the Market Rules assumed, a further slide in the deck stated: “[b]id to run at $● defines
76
Ibid. The PowerPoint presentation includes a second slide that is identical to the slide in Figure 7-2 except that the
CMSC payments for 2010 are shown as $0.95/MWh instead of $10.00/MWh. CMSC payments of $0.95/MWh are
consistent with the amount of CMSC payments received by Bowater when it was previously dispatchable (July 2003
to August 2006) and as calculated in a Bowater spreadsheet entitled “Power $/MWh 2010 Budget”, which shows
monthly CMSC receipts of between $0.92/MWh and $1.08/MWh for the years 2005 through 2008 (Responses to
RFI, B.3.3). Bowater also stated that “[p]reliminary estimate (sic) of CMSC revenue in the power cost reduction
initiative (sic) based on conservative historical amounts” (Responses to RFI, B.3). Bowater further stated that it
estimated potential CMSC revenue using a bid of $●/MWh and that “2010 Budget presentations were adjusted to
reflect this higher amount of CMSC revenues” (Responses to RFI, B.3).
77
Attachment to email from [Senior Bowater Personnel #2] to [Senior Bowater Personnel #5] and [Senior Bowater
Personnel #3], September 28, 2009. Responses RFI, B.3.6. Relevant excerpts are reproduced in Appendix J. The
concept of Marginal Benefit of Consumption is discussed in Section 6.3.2.2.
PUBLIC VERSION
Products Inc.
58
the [CMSC] compensation.”78 A detailed calculation in the presentation shows CMSC payments
of $38,005 during a ramp-down scenario based on the use of a bid price of $●/MWh.79 It is clear
that Bowater’s CMSC projections involved increases in operating profits rather than recovery of
operating profit reductions arising from responding to dispatch instructions caused by Grid
Conditions.
Personnel at Bowater and its affiliates were also aware that the IESO might seek to recover some
of the CMSC payments obtained as a result of their ramping strategies. As a result, on the advice
of Abitibi personnel, Bowater did not “book” the entire amount of CMSC payments that
appeared on its daily or monthly settlement statements from the IESO.80 Instead, Bowater
booked what it considered to be “legitimate CMSCs” arising from “planned shutdowns and startups… The balance was provisioned.”81 ABI was aware of the reasons for and supported this
accounting technique. For example, in March 2010, [Senior AbitibiBowater Inc Personnel #1]
enthusiastically reported to [Senior Bowater Personnel #3] and [Senior Bowater Personnel #5]
(copying [AbitibiBowater Inc Personnel #4]) that the CMSC amount included in the settlement
statement for February 2010 was substantially more than what originally had been booked:
78
Ibid.
Ibid.
80
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #3], February 23 and 24, 2010. Responses
to RFI, B.16.37.
81
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], March 15, 2010. Responses to RFI,
B.16.45.
79
PUBLIC VERSION
Products Inc.
59
From:
[Senior AbitibiBowater Inc Personnel #1]
To:
[Senior Bowater Personnel #3], [Senior Bowater
Personnel #5]
c.c.:
[AbitibiBowater Inc Personnel #4]
Date:
March 12, 2010 1:39 PM
Subject:
Power Bill Feb 2010
We received the bill. Our CMSC credits after clawbacks was
$1,405,008.74! Which is $623,188.74 higher than what we booked.
This is great news!82
In April 2010, Bowater increased the portion of the CMSC payments that it was booking after
monitoring the amounts being clawed back by the IESO. Bowater briefed ABI executives on its
plans in this regard:
From:
To:
[AbitibiBowater Inc Executive #1], [AbitibiBowater Inc
Executive #3]
c.c.:
Date:
April 22, 2010 11:06 AM
Subject:
Fw: March Electricity summary – Confidential
Please find below an explanation for our March electricity price.
Note the various programs we participate in from the different
sections of the Government. The March bill arrived last Friday,
April 16, and everything was as expected. For the month of April
82
PUBLIC VERSION
Products Inc.
60
we will increase our CMSC booking to 85%, and adjust the March
and February amounts accordingly. 83 (emphasis added)
In early June 2010, ABI financial executives decided to cease the practice of “holding
back” some CMSC and to “book” the full amount of CMSC received. In response to an
email containing a spreadsheet of the received and booked CMSC payments,
[AbitibiBowater Inc Executive #1] instructed the [AbitibiBowater Inc Personnel #4] as
follows:
From:
[AbitibiBowater Inc Executive #1]
To:
Date:
June 7, 2010 01:46 PM
Subject:
Re: TB Power – May 2010
Based on the conversation I had with [Senior Abitibi Personnel #2]
last week, we will need to book all of this in June and then every
month going forward, we will no longer keep any hold backs with
the exception of the amounts that [Senior Abitibi Personnel #2]
calculates as not true numbers.84 (emphasis added)
When pressed by the MAU on the rationale for Bowater’s high bid prices and the financial
impact if it was forced to reduce electricity consumption as a result of being constrained off
[Senior Bowater Personnel #3] alerted [Senior Abitibi Personnel #2] as well as [Senior
AbitibiBowater Inc Personnel #2] that there was a risk that Bowater’s CMSC payments were at
risk of being reduced or eliminated as shown in the following excerpt:
83
84
Responses to RFI, B.16.75. This email was subsequently forwarded to [Senior Bowater Personnel #3].
PUBLIC VERSION
Products Inc.
61
From:
To:
Cc:
Date:
July 29, 2010 06:06 PM
Subject:
Re: Fw: RE: Considerations relating to CMSC
Repayments by Abitibi
Interesting e-mail. Does he want to debate line item, by line item?
Decide for us what is valid and not? Does the market rules give
him the authority to ask? Confidentiality of our numbers if he gets
it.
Two schools of thought
1)
We comply, he reads and accepts our numbers, no rule
changes and we live happy ever after.
2)
[We] don’t comply, they go back to whoever and demand
a rule change to protect the consumer against those big bad loads!
End of CMSC from ramping sometime in the future?
3)
We comply, they don’t like our numbers and now [we’re]
negotiating against ourselves again. 85 (emphasis added)
These and other emails confirm that knowledge of the ramping strategy and its financial
consequences was not limited to operational employees at the Thunder Bay Facility, but
extended to senior management of Bowater and ABI.
Finding #2 (CMSC Ramping Strategy):
Bowater developed strategies to self-induce CMSC payments at the Thunder
Bay Facility, and these were known to senior management.
The main behaviours that triggered constrained-off CMSC payments for Bowater are analyzed in
the sections below.
85
PUBLIC VERSION
Products Inc.
62
7.4.2 Expanding the Magnitude of CMSC Using a High Bid Price
This section examines Bowater’s knowledge of the operating profit principles underlying the
CMSC regime (Section 7.4.2.1), and the relationship between its Marginal Benefit of
Consumption and bid prices (Sections 7.4.2.2 to7.4.2.5). The Panel has also analyzed the three
explanations for high bid prices that were provided by Bowater: 86
·
Bidding at a very high price reduced the risk of the facility being dispatched down
(which can occur if the Nodal Price is above the bid price) (see Section 7.4.2.6).
·
Bidding at a very high price reduced the risk of being activated to provide
operating reserve (while still being able to obtain revenue from participating in
the operating reserve market) (see Section 7.4.2.7).
·
Dispatchable loads owned by Abitibi-Consolidated at Fort Frances, Fort William,
and Iroquois Falls had bid at a similarly high price for a number of years (see
Section 7.4.2.8).
7.4.2.1 Bowater Understood the Operating Profit Principles in the CMSC Regime
Bowater and affiliated company personnel understood that the CMSC regime was designed to
compensate for reductions in operating profits based on an assumption that a dispatchable load’s
bids would reflect its Marginal Benefit of Consumption. In its initial discussions with the MAU
in June of 2010, Bowater defended its $●/MWh bid price citing the reasons stated above.
However, Bowater later acknowledged that $●/MWh did not reflect the financial impact of
incremental decreases or increases in consumption87 and lowered its bids to $●/MWh for
ramping hours.88 The internal communications related to this change included efforts to confirm
that the IESO training materials had not discussed the relationship between bids and the actual
financial impact of reduced consumption as shown in the following excerpt:
86
Responses to RFI, B.3, p. 2 and B13, p.1 and 2.
88
87
PUBLIC VERSION
Products Inc.
63
From:
To:
[Senior Abitibi Personnel #2], [Senior Bowater
Personnel #5], [Senior Abitibi Personnel #4]
Dated:
June 28, 2010 08:57 PM
Subject:
Fw: Considerations relating to CMSC Repayments
by Abitibi
[Senior Bowater Personnel #5] review your course notes/hand
written notes and ensure that during the training in January 2010
there was no mention of basing our bid on the “marginal lost
opportunity cost”. We were only trained on bidding high so as not
to be activated [for operating reserve].89
In fact, Bowater and Abitibi personnel knew that CMSC payments were intended to compensate
for operating profit reductions resulting from being dispatched differently than the economics of
the bids in the market schedule. As noted above, a presentation to senior management, dated
October 1st, 2009, contained a slide on CMSC which explained the operating profit calculation
and the assumption that dispatchable loads’ bids would reflect Marginal Benefit of
Consumption.90
Other documents submitted by Bowater in response to the Panel’s requests for information also
demonstrate awareness of the economic purpose of CMSC payments before and while the
Thunder Bay Facility was submitting a $●/MWh bid during its ramping hours. For example,
Bowater’s “March 2010 Purchased Electricity Summary”, prepared by [Senior Bowater
Personnel #5], included the following description of CMSC:
Congestion Management Settlement Credits (CMSC) – Paid by the
IESO due to being dispatchable. Credits are based [on] the
difference between the constrained and unconstrained dispatch
algorithms created by system constraints or the up and down ramp
rates of generators and loads. In our case when the TMP plant is
constrained to a different dispatch [than] what it had bid for, the
89
Responses to RFI, B.13.32. (The IESO training materials in relation to operating reserve activation are discussed
in Section 7.4.2.7 below.)
90
See the second slide reproduced in Appendix J and quoted in Section 7.4.1.
PUBLIC VERSION
Products Inc.
64
IESO provides a credit due to the lost opportunity to use that
energy. Ramping up and down helps stabilize the grid. CMSC
[costs] are shared by all consumers through uplift charges. 91
(emphasis added)
Finding #3 (Knowledge of CMSC Compensation Principles):
Bowater was aware that the CMSC regime assumed that dispatchable loads
would bid based on their Marginal Benefit of Consumption and that CMSC
payments were designed to compensate a dispatchable load for operating
profit reductions when it was directed by the IESO to follow a dispatch
different from its market schedule.
7.4.2.2 Bid Prices Exceeded Marginal Benefit of Consumption on Weekdays
Bowater’s bid prices were consistently and significantly above its own estimates of the Marginal
Benefit of Consumption.
As noted in Section 5.2.1, the MAU contacted Bowater in June 2010 to inquire about the
anomalously high CMSC payments being made to the Thunder Bay Facility during ramping
periods and whether they were obtained in part as a result of bid prices that exceeded the
Marginal Benefit of Consumption. Bowater responded by calculating and implementing a bid
price of $●/MWh during ramping periods beginning July 1, 2010, which it explained as follows:
In response to your first question on the calculation of the
opportunity-based bid, the the $●/ MWh opportunity cost is based
on lost production, i.e. not running the TMP plant for a one hour
period. The calculation uses the facilities net of paper, minus the
direct costs associated to arrive at a contribution/tonne. The
calculation is based on the amount of paper tonnes that would not
have been manufactured during the one hour of lost production on
the TMP plant. In addition to this, fixed cost recovery for the time
we did not produce paper and unavoidable direct costs such as
91
Attachment to an email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], April 20, 2010.
PUBLIC VERSION
Products Inc.
65
electricity, chemicals and steam are included. Also added to the
calculation is the restart time for the paper machine as it was shut
down due to lost inventory (pulp) made by the TMP plant. The
result is the sum of all the above, divided by one hour of TMP
power consumption to arrive at $●/MWh.
With respect to your second question regarding dispatch risk,
during our discussions and email exchanges regarding the
opportunity based bid cost, we changed the bid to $●/MWh from
$●/MWh. This change reflects opportunity costs lost during
ramping and does not reflect dispatch risk. Once our internal backoffice tools have been reconfigured to allow for this new
combination of price/quantity pairs structure, we will revert back
to bidding $●/MWh during the non-ramping hours to cover our
risk aversion.92
Bowater’s rationale for this calculation was that the Thunder Bay Facility was “pulp limited” as
a result of its obligation under its DR2 Agreement to operate its TMP plant off-peak on
weekdays. Pulp from the TMP plant feeds the Facility’s paper machine and, given limited pulp
storage capacity, lost production at the TMP plant could lead to production losses on the paper
machine later in the week when pulp inventories are lowest.93 Bowater estimated the loss of one
hour of TMP production would result in 3 hours of lost paper production which, based on its
calculations of contribution as well as certain costs that were characterized as unavoidable,
equated to $●/MWh.94
[Senior Bowater Personnel #5] also developed a “value of electricity” calculation of $●/MWh
for weekdays.95 This amount differs from the $●/MWh number originally used to support the
$●/MWh bid price, mainly because it does not incorporate an additional hour of lost paper
production due to a restart of the paper machine. A restart of the paper machine would not be
necessary if the paper machine was able to draw on pulp storage, which even at its lowest point
92
Excerpt of email from [Senior Abitibi Personnel #2] to MAU, July 29, 2010. Response to RFI, B.13.59.
94
95
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], July 22, 2010. Responses to RFI,
B.13.44.
93
PUBLIC VERSION
Products Inc.
66
on Friday evenings normally had over 5 hours of TMP reserves.96 Therefore, incorporating the
paper machine restart overstated the expected impact on operating profit if the Facility was
constrained off.
As noted above, Bowater indicated that its calculations included elements of “fixed cost
recovery”.97 Fixed costs normally do not change in response to short term transitory changes in
electricity consumption. Including fixed costs in the Marginal Benefit of Consumption could
result in cases where a positive operating profit that could have contributed to offsetting fixed
costs is foregone. Accordingly, they should not be included in calculating the Marginal Benefit
of Consumption. Thus the Bowater calculations referenced above overstate the actual Marginal
Benefit of Consumption. However, since the amount of the overstatement is not readily
determinable, the Panel has used Bowater’s own calculations as a conservative basis for
analyzing whether Bowater was bidding at prices which exceeded its Marginal Benefit of
Consumption.
The Panel also notes that, even on weekdays, Bowater’s typical operating pattern allowed
opportunities to make up lost production every morning. The morning ramp downs normally
were completed in HE 6 or HE 7 (i.e., one or two hours before Bowater was required to be
operating at its reduced level for the 8 a.m. to 6 p.m. peak hours under the DR2 Agreement).
While it is not necessary for purposes of this Investigation to make a definitive finding on this
point, it appears that the impact of being constrained off in a morning ramp down hour (and
potentially also on the ramp up hour the evening before) would likely be deferred rather than
permanently lost production, and that the calculations provided by Bowater based on lost pulp
and paper production would overstate the effect that being constrained off would have on its
operating profits for ramp down hours.
96
Data supplied in Bowater’s Responses to RFI, B.3.1 shows that average TMP consumption between the hours of 7
pm and 7 am on weekdays was ● tonnes, or ● tonnes per hour. The data also shows TMP inventory was lowest at 7
pm on Friday evenings at ● tonnes, which at a consumption rate of ● tonnes per hour represents over 5 hours of
TMP inventory before the paper machine would have to be stopped and restarted.
97
Email from [Senior Abitibi Personnel #2] to MAU, July 29, 2010, reproduced above. Responses to RFI, B.13.59.
PUBLIC VERSION
Products Inc.
67
As indicated in Section 7.2.1, when operating as a dispatchable load from 2003 to 2006, the
Thunder Bay Facility used bids in the range of $●/MWh to $●/MWh for much of its load during
normal operations and ramping. This is generally consistent with the conclusion that bid prices
of $●/MWh and $●/MWh were well above Bowater’s Marginal Benefit of Consumption on
weekdays.
7.4.2.3 Bid Prices Exceeded Marginal Benefit of Consumption on Weekends
In its responses to the Panel’s requests for information, Bowater produced an internal email
which indicated that when the Thunder Bay Facility was not pulp limited, such as on weekends,
the impact on operating profits from being constrained off would be lower than $●/MWh:
One thing we need [in respect of bid price rationale] is a better
story for the weekend. [I’m] comfortable we have a good
explanation for M-F, but [we’re] weak for Saturday and Sunday.98
In a subsequent email, [Senior Bowater Personnel #5] noted: “since we are not pulp limited on
the weekends as shown by the inventory outages, a lower price could be justified.”99
When determining what Bowater would propose to the MAU as a price that reflected the impact
of being constrained off during a weekend, [Senior Bowater Personnel #3] suggested that
“[m]aybe the price should be HOEP, I don’t know? But $● will give us a nice dividing
number”.100 [Senior Bowater Personnel #3] also requested that personnel provide a number for
“how much CMSC we have made on weekends, since we started.” 101 [Senior Bowater Personnel
#5] replied:
98
Email from [Senior Bowater Personnel #3] to [Senior Bowater Personnel #5], [Senior AbitibiBowater Inc
Personnel #1] and [Senior Bowater Personnel #1], July 31, 2010. Responses to RFI, B.13.65.
99
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], August 16, 2010. Responses to RFI,
B.13.86.
100
Email from [Senior Bowater Personnel #3] to [Senior Bowater Personnel #5], August 18, 2010. Responses to
RFI, B.13.91.
101
Email from [Senior Bowater Personnel #3] to [Senior Bowater Personnel #5], August 18, 2010. Responses to
RFI, B.13.91.
PUBLIC VERSION
Products Inc.
68
From:
To:
Date:
August 18, 2010 08:21 AM
Subject:
Re: CMSC updated per July billing
Weekends and stat holidays have generated about $2.5MM [in
CMSC payments] to end of July. Not sure what you are trying to
do on the weekends with the $●. One line shutdowns will generate
minimal CMSC at any price. Lowering the price however does
acknowledge that we have different economics on weekends
because of the excess pulp capacity. We could calculate a number
using the same format as during the week without the impact on
the machine. In any case I think we should have a higher number
than $● so that it doesn’t make the $● look so high.102 (emphasis
added)
As noted in Section 7.4.2.2, if the impact of being constrained off results in deferred rather than
permanently lost production, then even Bowater’s lower $●/MWh estimate might overstate the
impact that being constrained off while ramping on a weekend would have.103 However, it is not
necessary for purposes of this Investigation to make a definitive finding on this point.
7.4.2.4 Bowater’s Marginal Benefit of Consumption is Lower in Self-Induced Ramping Hours
Bowater’s various calculations also overstate the operating profit reduction that would result
from being constrained off during self-induced ramp down hours because of an implicit
assumption that the whole hour would be affected.
Based on its submitted ramp rates, the Thunder Bay Facility takes five intervals to ramp down
from ● MW to ● MW. During these five intervals, being constrained off would not have any
negative impact on Bowater’s operating profits because (i) Bowater’s bids indicate that it no
longer wishes to consume, and (ii) the Facility cannot be dispatched down faster or further than
102
As noted above, [Senior Bowater Personnel #3] identified HOEP as another possible measure of the impact of
being constrained off on weekends. The average HOEP on weekends during the Relevant Period was $40.46/MWh,
well below the $●/MWh amount noted in the same email. Responses to RFI, B.13, 91, quoted above.
103
PUBLIC VERSION
Products Inc.
69
the ramp rates and bid quantities it has submitted. Thus Bowater is only exposed to seven
intervals of potential lost or delayed TMP production if it were dispatched down at the beginning
of a self-induced ramp down hour.
Bowater did not demonstrate that being constrained off during some or all of the first seven
intervals of a ramp down hour would result in an overall loss of production that could not be
made up by adjusting its operations at a later time in which it had not planned to be running at
full capacity. In such circumstances, there would be little or no reduction of operating profit
from being instructed to ramp down a bit early on one morning. However, even if the production
was permanently lost, a calculation based on an entire hour of lost energy consumption
overstates the actual operating profit reduction.
In the Responses to RFI, Bowater stated that “during hours in which the Thunder Bay facility is
deliberately altering [its] consumption level … 65% of the ramp down hour TMP pulp is still
being produced.”104
If the estimates provided by Bowater for the impact of lost production and the unavoidable costs
of reducing consumption were otherwise correct, the estimated operating profit impact of being
constrained off during a weekday ramp down would range from $0/MWh (if not constrained in
intervals 1-7) to a maximum (if constrained from interval 1 forward) of $●/MWh
($●/MWh*65% using the calculation of the Marginal Benefit of Consumption in Bowater’s
internal correspondence in July 2010105 and ignoring possible overstatements from deferred
versus lost production and the inclusion of fixed cost recovery amounts).106
104
Responses to RFI, B.4. The Panel examined the total MWh of electricity consumption during a ramp down hour
relative to a full hour of consumption. During a normal ramp down hour the Thunder Bay Facility consumes 73% of
a full hour of electricity consumption. (This calculation assumes ● MWh of consumption during a full hour of
consumption and ● MWh of consumption based on a ramp down profile from ● MW to ● MW using the ramp rates
in Table 7-2.) The fact that the TMP pulp ratio is slightly lower likely reflects some additional use of electricity to
effect the shut down properly.
105
B.13.71.
106
Based on the calculation that Bowater submitted to the MAU, the prorated amount would be $●/MWh ($●/MWh
* 65%). (Responses to RFI, B.4, p.1, discussed in Section 7.4.2.2).
PUBLIC VERSION
Products Inc.
70
The Panel previously noted that [Senior Bowater Personnel #3]’s $●/MWh estimate of the
financial impact if the Thunder Bay Facility was constrained off during a weekend might
overstate the actual impact (see Section 7.4.2.3), but uses it as a conservative approximation in
the absence of other more detailed data. If pulp and paper production was not permanently lost,
but merely delayed for a period of time during a weekend constrained-off event, then there
would be little or no reduction of operating profits. Even if it is assumed that there would be
permanently lost production (which appears to be questionable on weekends), the maximum
estimated reduction of operating profit during a weekend ramp down hour if constrained off in
interval 1 would be $●/MWh ($●/MWh*65%).
During ramp ups, there is a possibility of pulp and paper production being reduced if the Facility
is constrained off in whole or in part during the first five intervals when ramping would normally
occur, as well as in the remaining seven intervals when the Facility is seeking to consume at its
bid quantity. Again, it is unclear whether these situations would involve delayed production that
could be made up later (e.g. by delaying the start of the next ramp down), and therefore little, if
any, impact on operating profits, or whether there would be permanently lost production.
Bowater indicated that TMP pulp is being produced during 75% of a ramp up hour.107
Accordingly, even if there would be permanently lost production as a result of being constrained
off for an entire ramp up hour, based on Bowater’s own estimates above, the operating profit
reduction would be no more than $●/MWh ($●/MWh*75%) on weekdays and $●/MWh
($●/MWh*75%) on weekends.
7.4.2.5 Bid Price Increased CMSC Payments
In summary, Bowater’s Marginal Benefit of Consumption was no higher than $●/MWh and
$●/MWh during ramp downs on weekdays and weekends, respectively. On ramp ups, it was no
higher than $●/MWh and $●/MWh on weekdays and weekends, respectively. However, the
107
Responses to RFI, B. 4. The Panel examined the total MWh of electricity consumption during a ramp up hour
relative to a full hour of consumption. During a normal ramp up hour the Thunder Bay Facility consumes 76% of a
full hour of electricity consumption. (This calculation assumes ● MWh of consumption during a full hour of
consumption and ● MWh of consumption based on a ramp up profile from ● MW to ● MW using the ramp rates in
Table 7-2.) The amount is almost identical to the TMP pulp ratio provided by Bowater.
PUBLIC VERSION
Products Inc.
71
actual financial impact of being constrained off while ramping may have been as little as zero in
situations where the facility was not capacity constrained and could make up the lost production
in a subsequent hour (particularly on weekends).
An estimate of the amount by which Bowater’s high bid prices triggered CMSC payments in
excess of operating profit reductions during self-induced ramp hours is set out in Table 7-5. This
estimate is based on the following conservative assumptions: (i) there were always permanent
losses of mill production resulting from being constrained off (as opposed to deferred production
that could be recouped on subsequent hours or days when the Facility was not operating at
capacity); (ii) estimates of the Marginal Benefit of Consumption are based on the conservative
(maximum) estimates outlined above which ignore potential overstatements related to fixed
costs; and (iii) all the quantity differences between the constrained schedule and the greater of
the unconstrained schedule and actual consumption reflected Bowater responding to IESO
dispatch instructions caused by Grid Conditions (which, as noted in Sections 7.4.3 – 7.4.7, was
not, in fact, the case). The total CMSC impact of Bowater’s high bid prices is estimated at $10.3
million.
PUBLIC VERSION
Products Inc.
72
Table 7-5: Estimated Impact of Bowater’s High Bid Prices on
CMSC Payments for the Thunder Bay Facility
($/MWh, MWh and $000)
Weekdays
Ramp Up
Ramp Down
Feb-June July-Aug Feb-June July-Aug
Weekends
Ramp Up
Ramp Down
Feb-June July-Aug Feb-June July-Aug
Bid Price
($/MWh)
●
●
●
●
●
●
●
●
Estimated
Marginal
Benefit of
Consumption
($/MWh)
●
●
●
●
●
●
●
●
Difference
($/MWh)
1,589
390
1,644
445
1,924
725
1,934
735
Constrainedoff Quantity
(MWh)*
1,262
492
2,643
1,067
500
204
982
440
Total CMSC
$ 2,005
$ 191
$ 4,345
$ 475
Impact ($000)
*For intervals where Bowater received a net CMSC payment.
$ 962
$ 148
$ 1,899
Finding #4 (Operating Profit Impact of Being Constrained
Off):
a) During periods when Bowater was not operating the Thunder Bay
Facility at capacity, there would be virtually no reduction in operating
profits as a result of being constrained off during a ramping hour
because production could be made up in a subsequent hour.
b) Even in situations where the Thunder Bay Facility was capacity
constrained, Bowater’s bid prices between February and August 2010
substantially exceeded its Marginal Benefit of Consumption and the
reduction in operating profits that would result from the Thunder Bay
Facility being constrained off during ramping hours.
PUBLIC VERSION
$ 323
Products Inc.
73
c) Based on data provided by Bowater, the difference between Bowater’s
February - June 2010 bid price of $●/MWh and its Marginal Benefit of
Consumption when ramping down was at least $1,644/MWh on weekdays
and $1,934/MWh on weekends.
d) Based on data provided by Bowater, the difference between Bowater’s
February - June 2010 bid price of $●/MWh and its Marginal Benefit of
Consumption when ramping up was at least $1,589/MWh on weekdays
and $1,924/MWh on weekends.
e) Based on data provided by Bowater, the difference between Bowater’s
July - August 2010 bid price of $●/MWh and its Marginal Benefit of
Consumption when ramping down was at least $445/MWh on weekdays
and $735/MWh on weekends.
f) Based on data provided by Bowater, the difference between Bowater’s
July - August 2010 bid price of $●/MWh and its Marginal Benefit of
Consumption when ramping up was at least $390/MWh on weekdays and
$725/MWh on weekends.
g) Bowater’s high bid prices were used to obtain CMSC payments that more
than compensated Bowater for operating profit reductions by at least
$10.3 million.
7.4.2.6 The Risk of Being Constrained Off Did Not Justify Bowater’s Bid Prices
Bowater’s assertion that high bid prices were necessary to deal with the risk of being constrained
off while ramping is not consistent with (i) the actual consequences of being constrained off, (ii)
the availability of other options to eliminate such a risk, and (iii) the fact that the risk was in fact
remote.
The impact of being constrained off during a ramping hour can be summarized as follows:
PUBLIC VERSION
Products Inc.
74
·
During periods where the Thunder Bay Facility is producing pulp, the impact of
being dispatched down or off was that the plant’s pulp operations would likely
slow down or stop (and, if there was insufficient pulp in storage, then paper
production would likely be similarly affected). However, during a self-induced
ramp down, the facility’s dispatch instructions were based on its own bid
quantities that reflected its desire to shut down the pulp operations.
·
During the five intervals of ramp down, dispatch instructions were reducing the
Thunder Bay Facility’s output at the ramp rate it submitted. Even if the Nodal
Price rose above $●/MWh in these intervals, the facility would not be dispatched
off any faster. Thus Bowater was at no risk of being dispatched any differently
than it desired during the actual ramp down intervals.
·
For the intervals before the scheduled ramp down begins (i.e. the first seven
intervals in the ramping hour), the Thunder Bay Facility was at risk of being
constrained off and losing pulp and possibly paper production (i.e. shutting down
a bit sooner) if the Nodal Price exceeded its bid price.
·
During a self-induced ramp up, the Thunder Bay Facility was dispatched to
increase its consumption in accordance with the quantity bids and ramp rates that
it submitted. It is possible that the Facility could be dispatched below its desired
ramp up path if the Nodal Price rose above the bid price during the five ramp up
intervals. This would result in a delay in reaching the planned pulp and possibly
paper production levels. It could also be dispatched below its planned
consumption level in the remaining seven intervals.
The foregoing summary of the possible impact of lost production during a ramping hour ignores
the availability of CMSC payments. If the Thunder Bay Facility was constrained off during a
ramp, Bowater would receive a CMSC payment. This would offset any negative effect on
PUBLIC VERSION
Products Inc.
75
Bowater’s operating profits (unless it was using a bid price lower than its Marginal Benefit of
Consumption, which was not the case at any time during the Relevant Period).108
Alternatively, Bowater could have bid at $2,000 MWh and become non-dispatchable. This
would have eliminated any risk of being dispatched down in the early rather than later part of a
ramp down hour, or during a ramp up hour. The Responses to RFI contain handwritten notes
made by [Senior Abitibi Personnel #2] which indicate that [Senior Abitibi Personnel #2] was
aware of the ability to bid at $2,000/MWh and that doing so would result in the facility being
treated as non-dispatchable.109 However, [Senior Abitibi Personnel #2] also understood that
operating on a non-dispatchable basis would mean the load would no longer be eligible for
CMSC payments. This is further confirmation that Bowater’s assertion that a $●/MWh bid price
was necessary to avoid the risk of being constrained off is not credible.
Bowater’s assertion that there was a material risk of being constrained off during ramping hours
is also not credible. The Northwest zone in the Ontario wholesale electricity market often has an
oversupply of energy, resulting in low or even negative Nodal Prices much of the time. Nodal
Prices above $●/MWh (or even Bowater’s revised July to August 2010 ramping bid price of
$●/MWh) are extremely rare at any time of day (including during HE 6 and HE 19, which were
the normal hours used for weekday ramping by the Thunder Bay Facility, as well as during
weekends). Thus, a high bid price normally would not be necessary to avoid being constrained
off during either a planned ramp up or the first seven intervals in a planned ramp down hour.
To test Bowater’s claim that a high bid price was necessary to avoid being dispatched down,
the Panel conducted an analysis of Nodal Prices during Bowater’s self-induced ramp hours on
weekdays and weekends at the Thunder Bay Facility between February 8, 2010 and August 28,
2010. The Panel also conducted an analysis of Nodal Prices in HE 6 and HE 19 (the Facility’s
usual ramp down and ramp up hours, once it again became dispatchable) on weekdays and
108
In fact, for the reasons discussed in Section 7.4.2.2 to 7.4.2.5 above, $●/MWh, $●/MWh or even $●/MWh bid
prices exceed Bowater’s Marginal Benefit of Consumption.
109
Notes from a meeting of the IESO’s Stakeholder Advisory Committee on March 31, 2010 (which is discussed in
relation to other issues in Section 8.5 below). Responses to RFI, B.11.6.
PUBLIC VERSION
Products Inc.
76
weekends between January 2009 and December 2009. The analysis considered Bowater’s
original $●/MWh bid price, its revised $●/MWh bid price, and its alternative calculation of the
Marginal Benefit of Consumption of $●/MWh contained in internal correspondence in July
2010.110 During the Relevant Period, there were 8,820 five-minute intervals in self-induced ramp
hours (6,384 on weekdays and 2,436 on weekends) and during 2009 there were 8,760 intervals in
HE 6 and HE 19 (6,000 on weekdays and 2,760 on weekends). The results are shown in Table 76:
110
B.13.71.
PUBLIC VERSION
Products Inc.
77
Table 7-6: Likelihood of the Thunder Bay Facility Being Constrained Off
at Various Bid Prices During Self-Induced Ramping Hours
January to December 2009 and February to August 2010
($/MWh, number and % of intervals)
Period
January to
December
2009
February to
August 2010
Bid
Price
($/MWh)
Number of Intervals Constrained Off
●
●
●
●
●
HE 6
0
0
0
1
1
●
●
●
●
●
HE 6
0
0
19
39
59
●
●
●
●
●
Ramp Down
0
0
0
0
1
●
●
●
●
●
Ramp Down
0
0
19
39
59
Weekday
HE 19
6
6
22
27
27
Weekend
HE 19
0
0
21
17
20
Weekday
Ramp Up
0
5
6
16
19
Weekend
Ramp Up
0
0
46
69
75
Percent of
Intervals
Constrained Off
Total
6
6
22
28
28
0.100%
0.100%
0.367%
0.467%
0.467%
Total
1
1
40
56
79
0.036%
0.036%
1.449%
2.029%
2.862%
Total
0
5
6
16
20
0.000%
0.078%
0.094%
0.251%
0.313%
Total
0
0
65
108
134
0.000%
0.000%
2.668%
4.433%
5.501%
This analysis confirms that a high bid price of $●/MWh, or even $●/MWh or $●/MWh, was
almost never necessary to prevent the Thunder Bay Facility from being dispatched down when
ramping. Based on the outcomes in 2009, which was the most recent information that would
PUBLIC VERSION
Products Inc.
78
have been available to Bowater before it again became dispatchable, the probability of being
constrained off was remote. Moreover, the results throughout the Relevant Period confirm that
the probability was remote. Even if Bowater had bid at the estimated levels of Marginal Benefit
of Consumption discussed in Section 7.4.2.5, the likelihood of being constrained off during a
ramping hour would have been remote on weekdays and very low on weekends (and any
negative impact on Bowater’s operating profits would have been compensated by a CMSC
payment).
Finding #5 (Risk of Being Constrained Off):
The risk of being constrained off during self-induced ramping hours did not
justify Bowater’s use of a bid price of $●/MWh or $●/MWh, or any other
level above the Marginal Benefit of Consumption of the Thunder Bay
Facility.
7.4.2.7 The Risk of Operating Reserve Activations Did Not Justify Bowater’s Bid Prices
As a dispatchable facility, Bowater is eligible to participate in the IESO’s operating reserve (OR)
market. During the Relevant Period Bowater received over $174,000 in OR (or OR CMSC)
payments in exchange for its offers to reduce energy consumption if activated to provide OR.
OR is standby capacity to produce power (or, equivalently, reduce load) that the IESO can call
upon with short notice when an unexpected event on the grid creates a need to rebalance supply
and demand. Dispatchable resources that are scheduled to provide OR receive standby payments
in exchange for being prepared to respond (i.e. to reduce consumption, in the case of a
dispatchable load) in the event of a contingency. When a contingency event occurs, and unless a
reliability concern dictates otherwise, the IESO will activate the resource with the lowest energy
offer or bid price to provide the required amount of OR up to the quantity of OR for which it has
been scheduled.111
111
The least costly resource is identified based on energy market offers or bids, since the activated power (or load
reduction) must be paid a minimum of the facility’s offer (or bid) price.
PUBLIC VERSION
Products Inc.
79
Bowater claimed that its strategy of using high bid prices to reduce the risk of being activated for
OR (while also being able to generate revenue from the OR market) “was confirmed in on-site
training provided by the IESO”.112 In fact, the IESO training materials simply include an
example which illustrates how the IESO decides to activate a dispatchable resource for operating
reserve. In the example, a load bids for energy at $1,990/MWh while a generator offers energy at
$200/MWh. Both also offer OR, and when an OR activation is required, the IESO selects the
generator because it is the less expensive resource.113 While the IESO training materials use a
very high bid price for the load, this is merely an illustration using hypothetical prices (that fall
between the MMCP and negative MMCP). The illustration also relates to a period of regular
operation by a dispatchable load, not a self-induced ramp up or ramp down. The training
materials do not instruct dispatchable loads to bid at very high levels or contrary to their
Marginal Benefit of Consumption during normal operations or during ramping periods.
If the Thunder Bay Facility was activated for OR during a ramping hour, Bowater would be
compensated for its reduced energy consumption based on its bid price. Thus, as long as its bid
price was not lower than its Marginal Benefit of Consumption, there would be no reduction of
operating profit as a result of an OR activation during a self-induced ramping hour.114 This
demonstrates that Bowater’s claim about the necessity of using a very high bid price to prevent
OR activations is not credible.
If the risk of being constrained off was in fact regarded as serious, Bowater could have chosen
not to offer OR during self-induced ramping hours. Alternatively, as noted in Section 7.4.2.6
above, it could have bid $2,000/MWh (i.e. an extra $●/MWh) to become non-dispatchable in the
energy market and thereby avoid the OR activation risk. Presumably it chose not to do so
112
See IESO, Introduction to Ontario’s Physical Markets, online:
http://www.ieso.ca/imoweb/pubs/training/IntroOntarioPhysicalMarkets.pdf, p. 58. (This training document was
revised in October 2011, but the OR example cited in Bowater’s Responses to RFI, B.3, is unchanged.)
114
For the reasons discussed in Sections 7.4.2.2 to 7.4.2.5, a $●/MWh, $●/MWh, or even $●/MWh bid price would
exceed the Facility’s Marginal Benefit of Consumption and more than compensate Bowater for reduced operating
profit if it was constrained off.
113
PUBLIC VERSION
Products Inc.
80
because this would have eliminated the availability of CMSC payments and/or the OR standby
payments.
The impact of being activated for OR during a self-induced ramping hour is similar to the impact
of being constrained off (described previously in Section 7.4.2.6):
— There is a risk of an earlier-than-expected reduction of energy consumption, and
hence lost or delayed pulp and possibly paper production, if an OR activation
occurs during the first seven intervals of a ramp-down hour.
— There is a risk of delayed energy consumption, and hence lost or delayed pulp and
possibly paper production, if an OR activation occurs during the ramp-up hour.
In order to assess Bowater’s claim that it was necessary to bid at a high price to avoid OR
activation, the Panel examined the frequency of OR activations in the Northwest zone during the
hours in which the Thunder Bay Facility was ramping up or down. There were none during the
Relevant Period.115 The Panel also examined the number of OR activations in HE 6 and HE 19
(the usual ramp down and ramp up hours after Bowater again became dispatchable) in 2009.
There were no activations in HE 19 and only two activations in HE 6, totaling 8 MW of
operating reserve activation. Thus the use of a high bid price was not necessary to mitigate OR
activation risk based on the information that would have been available when Bowater developed
its dispatchable load bidding strategy. Indeed, a presentation to ABI senior management in
October 2009 showed projected OR revenues averaging $●/MWh while at the same time noting
the miniscule activation risk: “Experience less than one dispatch per year (FF [Fort Frances]:
twice in 4 years)”.116
115
OR activations are infrequent in the Northwest. Of the 84 days during the Relevant Period with an OR activation
in Ontario, there were only four days in which OR was activated in the Northwest. Of these, only one affected the
Thunder Bay Facility. However, it did not occur in a ramping hour.
116
See “Thunder Bay 2010 Power Cost – October 1st, 2009.” Responses to RFI, B.3.6. p.3, reproduced in Appendix
J (the Fort Frances Facility is in the same region as the Thunder Bay Facility and the OR activation risks at the two
Facilities are similar.)
PUBLIC VERSION
Products Inc.
81
Finding #6 (Risk of Being Activated for Operating Reserve):
The risk of being activated to provide operating reserve during self-induced
ramping hours did not justify Bowater’s use of an energy market bid price of
$●/MWh or $●/MWh, or any other level above the Marginal Benefit of
Consumption of the Thunder Bay Facility.
7.4.2.8 Historical Use of High Bid Prices by Affiliates Did Not Justify Bowater’s Bid Prices
Bowater’s argument that a $●/MWh bid price was used by the Abitibi Consolidated dispatchable
loads at Fort Frances, Fort William and Iroquois Falls117 is not a justification for Bowater’s high
bid prices for at least four reasons. First, the fact that other dispatchable loads may have engaged
in conduct that may be exploiting a market defect does not have any bearing on whether Bowater
was exploiting a market defect during the Relevant Period. Second, it is possible that other
facilities could have a Marginal Benefit of Consumption equal to or greater than this level,
whereas Bowater does not (see Sections 7.4.2.2 to 7.4.2.5). Third, it is possible that high bid
prices may be used by dispatchable loads that ramp quickly and therefore trigger negligible
CMSC payments. Fourth, the specific reference to Abitibi’s use of this bidding strategy carries
no weight based on the separate analysis of the bid prices used by the Fort Frances Facility in
Section 8.4.2 of this Report.
Finding #7 (High Bid Prices by Other Loads):
The historical use of high bid prices by other dispatchable loads does not
provide a justification for Bowater’s high bid prices during self-induced
ramping hours.
117
PUBLIC VERSION
Products Inc.
82
7.4.3 Submitting Maximum Bid Quantities in Excess of Consumption
When Bowater began to develop its ramping strategy in September 2009, [Senior Bowater
Personnel #5] outlined the Thunder Bay Facility’s normal operations in a PowerPoint
presentation. The presentation shows the facility consuming between ● MW and ● MW when
both TMP lines were in operation.118 In subsequent emails between personnel at the Thunder
Bay Facility and the Fort Frances Facility, the load at the Thunder Bay Facility is referred as
being in the ● MW to ● MW range.119
Shortly after Bowater re-entered the market as a dispatchable load, [Senior Abitibi Personnel #2]
advised that increasing the quantity bid from ● MW to ● MW would increase CMSC payments
by $4,000 per ramp-up. It is notable that the suggested change originated from [Senior Abitibi
Personnel #2] rather than the personnel responsible for operating the Thunder Bay Facility. The
appropriateness of the proposed change (as well as the appropriateness of ramping below the
facility’s dispatch signal, which is analyzed in Section 7.4.5 below) was explored by email.120
[Senior Bowater Personnel #5] concluded that “[t]he others need to hear your recommendations
and know that there is still money to be had here. How we go after it will be the question?”.121
Two days later [Senior Bowater Personnel #5] informed [Senior Abitibi Personnel #2] that
“[b]ids have been changed to ● MW vs ● MW as you recommended.”122
From February 19, 2010 to May 11, 2010 Bowater bid ● MW instead of ● MW as its maximum
bid quantity. Although Bowater consistently bid to consume a maximum bid quantity of ● MW,
118
Email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], September 11, 2009. Responses to
RFI, B.2.5.
119
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #5], September 11, 2009, Responses to
RFI, B.2.6; email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], September 24, 2009,
Responses to RFI, B.2.10; email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], February 8,
2010, Responses to RFI, B.2.14; and attachment to email from [Senior Bowater Personnel #5] to [Senior
AbitibiBowater Inc Personnel #2] and [Senior Abitibi Personnel #2], January 22, 2010, Responses to RFI, B.2.25.
120
Email exchange between [Senior Bowater Personnel #5] and [Senior Abitibi Personnel #2], February 12-16,
2010. Responses to RFI, B.2.16.
121
Email exchange between [Senior Bowater Personnel #5] and [Senior Abitibi Personnel #2], February 16, 2010.
122
Email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], February 18, 2010. Responses to
RFI, B.2.17.
PUBLIC VERSION
Products Inc.
83
the Thunder Bay Facility’s consumption varied and followed a decreasing trend during this
period. In the month of February and the first half of the month of March, the Thunder Bay
Facility’s consumption varied between ● MW and ● MW, and at times consumption was well
below ● MW. Beginning in the middle of March the Thunder Bay Facility’s consumption was
declining, and varied between ● MW and ● MW. By the time Bowater revised its bids down
from ● MW in May, the facility was consuming between ● MW and ● MW, well below the bid
quantity.
Personnel at the Thunder Bay Facility recognized that consumption was usually less than ● MW.
For example, [Senior Bowater Personnel #5] observed in March 2010 that the load was incapable
of achieving a target of ● MW on start-ups and on occasion was significantly below target prior
to shutdowns.123 In emails on May 4 and 5, 2010, [Senior Bowater Personnel #5] reported on the
actual performance of the Thunder Bay Facility with statistics showing consumption in the range
of ● MW to ● MW.124 Six days later, Bowater revised down its maximum bid quantity from
● MW to ● MW.125
The Market Rules allow for variations in consumption around the dispatch instruction (referred
to as the “Compliance Deadband”). The Compliance Deadband for a resource with the
characteristics of the Thunder Bay Facility is 15 MW above or below its dispatch instruction.126
When dispatched at ● MW, the Facility is considered compliant while consuming anywhere
between ● MW and ● MW. During the period when it used a maximum bid quantity of ● MW,
123
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], [Senior Bowater Personnel #1],
[Senior Bowater Personnel #2], [Senior Bowater Personnel #4] and [Senior Abitibi Personnel #2], March 22, 2010.
124
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #2], May 4, 2010, Responses to RFI,
B.2.19; email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], May 4, 2010, Responses to RFI,
B.2.20; and email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #2], May 5, 2010, Responses
to RFI, B.2.21.
125
On September 2, 2010, [Senior Bowater Personnel #5] noted the possibility of a further change in the bid
quantity as an option for reducing the high CMSC payments that had been identified by the MSP and IESO:
“Reduce target from ● to ● MW (closer to actual)” and “Lower target [of] ● MW for start up hour (closer to what
we actually achieve)” (emphasis added). Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel
#3], September 2, 2010. Responses to RFI, B.13.111 (reproduced in full below).
126
IESO, Market Rule Interpretation Bulletin, Compliance with Dispatch Instructions Issued to Dispatchable
Facilities, June 29, 2009, online: http://www.ieso.ca/imoweb/pubs/interpretBulletins/ib_IMO_MKRI_0001.pdf.
PUBLIC VERSION
Products Inc.
84
the Thunder Bay Facility was never consuming above the Compliance Deadband. However, it
consumed below its Compliance Deadband in 990 intervals. Had Bowater bid a maximum
quantity of ● MW, the facility would have consumed above the Compliance Deadband in only
1 interval and below in 300 intervals.
By raising its maximum bid quantity to ● MW, Bowater increased the number of constrained-off
MWs for which CMSC payments would be made. Compared to a bid quantity of ● MW,
Bowater received CMSC payments for an additional 5 MW in each interval in which it was
constrained off. After March 15, 2010 and until May 11, 2010, Bowater received a net CMSC
payment after clawbacks in 402 intervals during which it had an unconstrained schedule of ●
MW. As a result, Bowater was paid approximately $330,000 in additional constrained-off CMSC
payments based on its submitted quantity bid of ● MW, relative to the CMSC payments that
would have been made if a ● MW maximum bid level quantity had been used.127
Finding #8 (Maximum Bid Quantity):
a) Bowater’s change in its maximum bid quantity from ● MW to ● MW from
February 19 to May 11, 2010 was undertaken to, and did, increase
constrained-off CMSC payments.
b) The estimated amount of incremental CMSC payments derived from
Bowater’s use of a ● MW maximum bid quantity at the prices Bowater was
bidding was $330,000.
127
This calculation is based on Bowater’s actual bid prices. It therefore overlaps with the estimate of the impact of
Bowater’s high bid prices in Table 7-5. If Bowater had bid at the Panel’s conservative estimates of the Facility’s
Marginal Benefit of Consumption of $●/MWh during non-ramping hours, $●/MWh and $●/MWh during ramp up
and ramp down respectively on weekdays, and $●/MWh and $●/MWh during ramp up and ramp down respectively
on weekends (as estimated in Sections 7.4.2.2 and 7.4.2.23), the extra CMSC payments related to the elevated
maximum bid quantity would have amounted to approximately $42,000.
PUBLIC VERSION
Products Inc.
85
7.4.4 Expanding Schedule Quantity Differences Through Ramp Down Timing
Beginning in September 2009, Bowater was planning a ramping sequence for the Thunder Bay
Facility that would have the load ramp down by 7:00 a.m. (i.e. interval 12 in HE 7) every
weekday, and back up starting at 7:00 p.m. (i.e. interval 12 in HE 19) every weekday evening.128
Figure 7-3 shows the typical sequence of ramp down steps and dispatch instructions during the
Relevant Period.
Figure 7-3: Typical Ramp Down Pattern Used by the Thunder Bay Facility
(MW by interval)
Of note in Figure 7-3 is the low ramp rate attributable to the shutdown of the auxiliaries. Bidding
the auxiliaries to shut down in the same hour as the main line and rejects refiners increased the
quantity differences between the constrained and unconstrained schedules. To provide the
minimum commitment of ● MW of demand reduction under the DR2 Agreement, it was not
necessary to ramp the auxiliaries down in HE 7. In fact, [Senior Bowater Personnel #5] had
128
Bowater’s DR2 Agreement required a demand reduction of ● MW during on-peak hours, which were defined as
8 a.m. (i.e. interval 12 in HE 8) to 6 p.m. (i.e. interval 12 in HE 18).
PUBLIC VERSION
Products Inc.
86
recognized this flexibility with respect to ramping the auxiliaries down prior to re-registering as a
dispatchable load:
I have set the strategy up with DR2 in mind and our need to stay
out of the peak period of 7 am to 7 pm with any significant load,
but to also maximize the up time of the refiners in the off peak
period. As a result you will see that I have given the operation the
flexibility to run the auxiliaries in the on peak period. So on the
way down set [our] bids up to go to ● MW at 7am and to ● MW at
8am, however we would shut the auxiliaries down anytime
between 7am and 8 am, probably within 15 minutes.129
Despite bidding the auxiliaries down in HE 7, personnel at the Thunder Bay Facility were
nonetheless instructed that it was permissible to ramp auxiliaries down into the start of the next
hour.130 The Panel also notes the contradictory explanations provided by Bowater for the
operations of the Facility. On the one hand, Bowater claimed that it submitted a high bid price of
$●/MWh (or $●/MWh) in a ramp hour to avoid being dispatched down and to have a high
uptime of the TMP in the off-peak hours in order to maximize the payments under its DR2
Agreement with the OPA.131 On the other hand, Bowater elected to bid its auxiliaries down in
HE 7, which reduced the uptime of the TMP in the off-peak hours and contradicted the analysis
provided by [Senior Bowater Personnel #5].
Had Bowater bid to ramp the auxiliaries down in HE 8 rather than in HE 7, the CMSC payments
during a ramp down would have been much less, as shown in Figure 7-4. In addition, the Facility
would have been able to consume electricity (and produce pulp and paper) for two additional
intervals in HE 7 at market prices well below its bid price. The Thunder Bay Facility could have
maintained an orderly shutdown of its pulping operations by ramping its mainline and reject lines
during HE 7 and its auxiliaries during HE 8. The sequence of the shutdown would remain
unchanged (see the actual curve in Figure 7-3) but the CMSC payments would be significantly
129
Email from [Senior Bowater Personnel #5] to [Senior Abitibi Personnel #2], September 24, 2009. Responses to
RFI, B.2.10.
130
Email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #4], February 1, 2010. Responses to
RFI, B.2.28.
131
Responses to RFI, B.2, p.1 and Responses to RFI, B.3.
PUBLIC VERSION
Products Inc.
87
reduced. When Bowater was using a $●/MWh bid price, the CMSC payments under the ramping
scenario in Figure 7-4 would have been approximately $30,000,132 compared to payments of
over $50,000 on a typical ramp down using the approach adopted by Bowater.
Figure 7-4: Alternative Ramp Down Bid Structure for the Thunder Bay Facility
(MW)
Bowater asserted that its ramp rates and sequence were necessary to ensure an orderly shutdown
and start up of the plant. When the Thunder Bay Facility first submitted its ramp rates the week
before it became dispatchable, IESO staff noticed that they were very low and contacted the
Facility.133 Bowater explained that the ramp rates were designed “to have an orderly shutdown
and have the TMP plant completely down by 7 am” and that the Facility required time “to purge
the process, minimizing the potential plugged lines and equipment that would delay start up.”134
132
This calculation is based on an average MCP of $30/MWh.
134
133
PUBLIC VERSION
Products Inc.
88
However, after an urgent Market Rule amendment was introduced by the IESO on August 28,
2010 to temporarily eliminate all constrained-off CMSC payments to dispatchable loads,135
personnel at the Thunder Bay Facility contemplated a number of changes to future bids that
would reduce CMSC payments — including a new ramping sequence whereby it would ramp the
auxiliaries down in the hour after, instead of the same hour as, the refiners were ramped down.
[Senior Bowater Personnel #5] reported to the [Senior Bowater Personnel #3] as follows:
From:
To:
Cc:
[Senior Bowater Personnel #1], [Senior Bowater
Personnel #2]
Date:
September 2, 2010 11:11 AM
Subject:
Potential Bid Change and impact
Changing the bids as follows:
Reduce target from ● to ● MW (closer to actual)
Bid auxiliaries down in second hour instead of the same hour
(reverse of start up)
Lower target or ● MW for start up hour (closer to what we actually
achieve)
One line shutdowns for inventory management on weekends
CMSC
-$10,200 for shutdowns
-$6,200 per startup
Monthly total $328,000 based on 20 days. No CMSC generated on
weekends. If ramp downs are totally eliminated with pending rule
135
IESO, Urgent Rule Amendment Proposal, MR-00373-R00, August 27, 2010, online:
http://www.ieso.ca/Documents/mr/MR_00373-R00.pdf. (This change is discussed in Section 9.)
PUBLIC VERSION
Products Inc.
89
change and only ramp ups are left monthly CMSC will be about
$124,000 per month with good start ups.136 (emphasis added)
The evidence regarding potential changes to operations after the Urgent Market Rule
Amendment (after the Relevant Period) is consistent with the evidence regarding actual
operations prior to Bowater’s re-registration as a dispatchable load (before the Relevant Period).
Both demonstrate that during the Relevant Period the Thunder Bay Facility was operated using a
ramping sequence that was not operationally necessary, and that had the effect of significantly
increasing CMSC payments.
Table 7-7 contains an estimate of the incremental impact of the alternative ramp down timing
identified in the above email on CMSC payments. Bowater’s actual ramping timing generated
an additional $3.9 million in CMSC payments when measured at Bowater’s actual bid prices
(i.e., $8.0 million under Bowater’s actual ramping pattern compared to $4.1 million under the
alternative ramping pattern). If Bowater had been bidding at its Marginal Benefit of
Consumption (based on the Panel’s conservative maximum estimates set out in Section 7.4.2.4),
the differential between the two ramping patterns would have been $0.6 million in CMSC
payments ($1.2 million under Bowater’s actual ramping pattern compared with $0.6 million
under the alternative ramping pattern).137
136
To avoid overlapping calculations with section 7.4.2 and Finding #4, the estimate of the incremental CMSC
impact from Bowater’s ramp down pattern is calculated based on the Panel’s conservative estimate of Bowater’s
maximum Marginal Benefit of Consumption.
137
PUBLIC VERSION
Products Inc.
90
Table 7-7: Estimated Impact of Bowater’s Ramp Down Timing
on CMSC Payments for the Thunder Bay Facility
($/MWh, MW and $000)
Alternative Ramp Down Timing
Bowater’s Ramp Down Timing
@Bowater’s Bid
Prices
@Marginal Benefit of
Consumption**
@Bowater’s Bid
Prices
@Marginal Benefit of
Consumption**
Feb-June
Jul-Aug
Weekday
Weekend
Feb-June
July-Aug
Weekday
Weekend
Bid Price
($/MWh)
●
●
●
●
●
●
●
●
Average MCP
($/MWh)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
$/Ramp Down
57,043
22,127
9,338
841
29,505
11,445
4,830
435
Ramp Downs
(#)*
120
53
127
46
120
53
127
46
Total CMSC
($000)
6,845
1,173
1,186
39
3,541
607
613
20
Price
Difference
($/MWh)
Quantity
Difference
(Constrainedoff
MWh/Ramp)
$8,018
$1,225
$4,147
$633
** Based on the estimated maximum Marginal Benefit of Consumption as set out in Section 7.4.2.4 and Finding #4.
* Based on the number of ramp downs with a bid quantity change of ● MW or more (indicating a shutdown of both
TMP lines, rejects refiners and auxiliaries).
PUBLIC VERSION
Products Inc.
91
Finding #9 (Ramp Down Timing):
a) Bowater used a ramp down pattern for its auxiliaries that triggered
increased CMSC payments during the Relevant Period when there was a
known alternative ramping pattern that would have generated
significantly lower CMSC payments and that was compatible with the
Thunder Bay Facility’s operational requirements (having been used
before and considered for use after the Relevant Period).
b) The estimated amount of incremental CMSC payments derived from
Bowater’s ramp down pattern was $3.9 million at the prices Bowater was
bidding.
7.4.5 Ramping Faster than Submitted Ramp Rates
Ramping down at faster than submitted ramp rates is significant for two reasons. First, it
increases the difference between the quantities (constrained schedules or actual consumption,
versus unconstrained schedules) used to calculate CMSC payments.138 In addition, it indicates
that the submitted ramp rates were lower than the facility’s capabilities and operating levels,
which results in a longer ramping period and higher CMSC payments.
An analysis of the ramping behaviour of the Thunder Bay Facility reveals that the Facility
ramped down faster than its submitted ramp rates in at least one interval during 82% (191 of
233) of its ramp downs during the Relevant Period.
138
Ramping faster than the submitted rate has an impact because of the way in which CMSC payments are
calculated. The value for the dispatch quantity is derived from the constrained schedule at the end of each fiveminute interval. The value for actual consumption is derived from the revenue meter data, which is averaged over
the five-minute interval. As indicated in Appendix F, a facility that follows dispatch instructions which match its
ramp rates on a ramp down will meet the dispatch consumption target at the end of the interval and will therefore
have average consumption greater than this amount over the interval. Ramping down faster than submitted ramp
rates will bring the average of the revenue meter data over a five-minute interval closer to what the dispatch
schedule indicates at the end of the five-minute interval.
PUBLIC VERSION
Products Inc.
92
Had Bowater submitted the faster ramp rates it actually followed, the dispatch algorithm would
have created a dispatch schedule that was closer to the market schedule, resulting in smaller
quantity differences and lower CMSC payments.
Bowater personnel were aware of this aspect of the CMSC formula. For example, [Senior
Bowater Personnel #5] explained the concept to [Senior Bowater Personnel #3] using the chart in
Appendix L. Similarly, shortly after the Thunder Bay Facility again became dispatchable,
personnel from the Fort Frances Facility and the Thunder Bay Facility exchanged multiple
emails regarding the review of ramp down events and CMSC payments. Their awareness of the
impact of ramping faster than submitted rates is indicated in exchanges such as the following:
From:
To:
Date:
February 12, 2010 1:45PM
Subject:
Re: Fw: Thunder Bay Bowater Bids /Offers in the Trade App
In order for us to be closer to the dispatch eng amount we would
have to anticipate the dispatch signal and take the chips off earlier.
Danger is that the load may come off before the dispatch. It would
also show that we are capable of a much faster down ramp rate.
The first constrained dispatch signal goes from ● MW to ● MW
with the expectation that we will be at ● MW by the end of the 5
minutes, which we are.... Your points are well taken, however with
the CMSC potential I wonder if we shouldn’t try and stay a bit
under the radar as well as be able to defend the way we operate if
questioned.139 (emphasis added)
A few days later, [Senior Bowater Personnel #5] advised [Senior Abitibi Personnel #2] at the
Fort Frances Facility as follows:
As far as following the dispatch, we are meeting the requirements
of the dispatch if we are at or below the dispatch energy level by
the end of the interval, not averaging that amount. The only way to
139
Responses to RFI, B.2.15. There were also further email exchanges between [Senior Abitibi Personnel #2] and
[Senior Bowater Personnel #5] to similar effect on February 8 and 12, 2010: Responses to RFI, B.2.14, B.2.15 and
B.2.16.
PUBLIC VERSION
Products Inc.
93
actually average less than or equal to the dispatch energy amount is
start before the dispatch or remove more load and get far enough
below the dispatch energy amount sufficiently before the end of
the interval so that [the] average of the energy consumed in that
5 minute interval is less [then] dispatch amount. We don’t want to
go [too] far below the dispatch level or they may think that we are
too conservative with our ramp rates and can shutdown faster. This
is turning out to be a significant amount of CMSC payments, close
to 10% of the amount paid out to the whole province in 2008,
generators included. I can’t believe that IESO won’t be taking a
close look at this. What we are doing now is very defendable, not
sure I can say that if we start before the dispatch [signal], as it is
not necessary to meet the target energy level by the end of the
interval.140 (emphasis added)
An excerpt from a subsequent email confirms that [Senior Bowater Personnel #5] and [Senior
Abitibi Personnel #2] were aware that ramping down faster than the dispatch schedule could
constitute gaming:
From:
To:
Date:
February 16, 2010 10:16AM
Subject:
Re: Fw: Thunder Bay Bowater Bids /Offers in the
Trade App
Wouldn’t anticipating the dispatch be considered gaming.
There is also potential that the [load will] come off before the
dispatch signal is received. If we were bidding economically we
would not be able to anticipate the dispatch when dispatched off on
price[.] Having to allow time for removing the feed to the refiners
is part of our justification for the ramp rates that we have. There
are more subtle moves we can make prior to our dispatch that will
reduce our transport lag time as well as move us to [an] energy
140
RFI, B.2.16.
PUBLIC VERSION
Products Inc.
94
level below the energy dispatch target getting us closer on average
to the constrained dispatch energy target.141 (emphasis added)
The internal emails indicate that Bowater was attempting to increase its CMSC payments on
ramp down by getting its revenue meter data closer to or below the dispatch schedule. This is
also evident in Bowater’s analyses of shutdown and start-up scenarios for the TMP plant. For
example, attached as Appendix K is a spreadsheet prepared by Thunder Bay Facility personnel
which shows how they planned and forecasted their “Optimized CMSC Payment (actual load <
constrained schedule).”142 As the title indicates, the objective was to achieve actual load
consumption that was less than the dispatch schedule.143
Bowater also acknowledged that it could ramp faster than its submitted ramp rates in a
presentation on reducing power costs prepared for creditors in November 2009 (when Bowater
was in the midst of formulating its ramping strategy):
The Congestion Management credit ($10.00) is a side benefit from
participating in the OR market and from shutting down and starting
up every week day for DR2. Since we would much rather shut
down as quickly as possible but the grid operator request us to
ramp down to protect the integrity of the grid, we get compensated.
Conservative number based on FF experience.144 (emphasis
added)
Bowater’s comment that the IESO instructed it to ramp down slower than it preferred does not
reflect IESO practice. The IESO does not normally instruct market participants on what ramp
rates to submit and Bowater provided no evidence to the contrary.
141
Email attachment from [Senior Bowater Personnel #5] to [Senior AbitibiBowater Inc Personnel #2], [Senior
Abitibi Personnel #2] and [Senior Bowater Personnel #3], January 22, 2010. Responses to RFI, B.2.25.
143
CMSC payments are made for the difference between the unconstrained schedule and the greater of actual
consumption or the constrained schedule – see Appendix F.
144
PowerPoint attached to email from [Senior Bowater Personnel #3] to [AbitibiBowater Inc Executive #3],
November 12, 2009. Responses to RFI, B.16.25. In fact, CMSC payments result primarily from participation as a
dispatchable facility in the energy market. (CMSC payments may also result from participation in the OR market,
such as when the facility is constrained on to provide OR during OR shortages, or when the facility is constrained
off during OR activations).
142
PUBLIC VERSION
Products Inc.
95
An example of one of the 191 ramp downs where the Thunder Bay Facility ramped faster than
submitted ramp rates is June 7, 2010 in HE 6. Bowater bid $●/MWh and submitted the ramp
rates in Table 7-2. The ramping of consumption, and the CMSC payments triggered during the
ramping intervals, are shown in Table 7-8. The actual CMSC payments earned during the ramp
down ($54,933) exceeded the CMSC payments that would have been made had the Thunder Bay
Facility ramped from one dispatch instruction to the next at its submitted ramp rate ($50,712).
The Facility ramped down faster than its submitted ramp rates in intervals 10 and 11. In interval
10 it ramped from ● MW to ● MW in a five-minute period. According to Bowater’s submitted
ramp rates, the Facility was only capable of reaching ● MW from a starting point of ● MW, not
● MW. Similarly, in interval 11, the Facility was only capable of reaching ● MW from a starting
point of ● MW, not ● MW.145 Although the deviations appear small, each increase in constrained
off MWs was paid the difference between Bowater’s bid price of $●/MWh and the MCP (which
was approximately $33/MWh during these intervals).
145
Bowater’s actual ramp rate in interval 11 of ● MW/minute was 57% faster than its submitted ramp rate of ●
MW/minute.
PUBLIC VERSION
Products Inc.
96
Table 7-8: CMSC Payments on a Fast Ramp Down of the Thunder Bay Facility
June 7, 2010, HE 6
(MW, $/MWh and $)
Interval
1
2
3
4
5
6
7
8
9
10
11
12
Total
Unconstrained
Schedule
Constrained
Schedule
Actual
Consumption
MCP
Net
CMSC
Expected
Consumption
Expected
Net CMSC
(MW)
(MW)
(MW)
($/MWh)
17.15
17.60
17.95
17.95
26.78
28.57
30.00
31.20
31.58
32.54
33.35
33.53
($)
(MW)
($)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
4,630
9,844
12,532
13,760
14,168
$54,933
●
●
●
●
●
●
●
●
●
●
●
●
2,870
8,607
12,045
13,186
14,004
$50,712
The Panel has not estimated the aggregate incremental impact of Bowater’s faster ramp down on
CMSC payments as it is partially subsumed in the estimate contained in Section 7.4.2.5. This
calculation is partially subsumed in Section 7.4.2.5 because the estimate in that section is based
on the difference between the unconstrained and the greater of the constrained schedule and the
actual quantity consumed which accounts for any constrained-off megawatts from fast ramping.
The estimate in Section 7.4.2.5, however, only accounts for the incremental CMSC payments for
fast ramping constrained-off megawatts based on the difference between Bowater’s actual bid
prices and the estimated Marginal Benefit of Consumption of the Thunder Bay Facility. Fast
ramping constrained-off megawatts are entirely self-induced and should not be compensated for
at any bid price. The estimate in Section 7.4.2.5 therefore underestimates the impact of fast
ramping on constrained-off CMSC payments by the number of fast ramping constrained-off
megawatts multiplied by the difference between the estimated Marginal Benefit of Consumption
of the Thunder Bay Facility and the MCP. The Panel has not undertaken an interval by interval
estimation of the incremental impact of Bowater’s fast ramping beyond what is already
PUBLIC VERSION
Products Inc.
97
accounted for in Section 7.4.2.5. The Panel is nevertheless satisfied that the vast majority of
CMSC payments associated with Bowater’s fast ramping is subsumed in Section 7.4.2.5.
Finding #10 (Ramping Down Faster than Submitted Rates):
a) Bowater’s Thunder Bay Facility was able to, and frequently did, ramp
down faster than its submitted ramp rates during the Relevant Period,
indicating that its submitted ramp rates were lower than the Facility’s
operational capabilities.
b) The submission of ramp down rates that were lower than the Facility’s
operational capabilities increased the magnitude of constrained-off CMSC
payments to Bowater.
c) The ramping down of the Facility faster than the submitted ramp rates
increased the magnitude of constrained-off CMSC payments to Bowater.
7.4.6 Failure to Ramp
During the course of the Investigation, it was noted that during the Relevant Period the Thunder
Bay Facility occasionally failed to ramp up and/or down, even though Bowater had submitted
bid quantities that indicated it wanted to increase or decrease energy consumption. Such failures
to ramp resulted in differences between the market and dispatch schedules, and therefore
triggered constrained-off CMSC payments.
For example, on April 11, 2010, in HE 9 and HE 12, Bowater twice bid to ramp up, but then
failed to follow the dispatch schedule that reflected the planned ramp. The scheduled and actual
consumption, as well as the CMSC payments triggered during the planned ramping periods, are
illustrated in Figure 7-5:
PUBLIC VERSION
Products Inc.
98
Figure 7-5: Scheduled and Actual Consumption and CMSC Payments
at the Thunder Bay Facility During a Failure to Ramp
April 11, 2010
(MW and $ Dollars)
The market schedule was determined by the Facility’s submitted bid quantity, which was ● MW
for HE 9 and HE 12. The dispatch schedule was determined, in part, by the Facility’s actual
consumption in the prior interval, since it can only be moved within a “Dispatch Envelope” (the
range between the maximum and minimum dispatch instruction based on the load’s current
consumption level and submitted ramp rates). Because the Facility did not increase its energy
consumption, the dispatch schedule remained at around ● MW. Had the Facility followed its
dispatch instructions: (i) the dispatch schedule would have increased toward the market
schedule, (ii) there would have been smaller quantity differences between the two schedules, and
(iii) smaller CMSC payments would have been triggered. By not increasing its consumption
levels (i.e. failing to ramp), Bowater received over $200,000 in net CMSC payments (the highest
daily CMSC payment made to the Thunder Bay Facility during the Relevant Period).
PUBLIC VERSION
Products Inc.
99
Bowater stated that this and other failures to ramp occurred because the Facility experienced
equipment failures.146 Failure to ramp occurred on an infrequent basis. In the absence of
evidence indicating that these failures to ramp were deliberate, the Panel has concluded that this
behaviour was not an attempt to exploit a market defect. Even though the CMSC payments
triggered by the failures to ramp did not arise from exploitative conduct, they were unwarranted
and should have been clawed back because they were caused by the conditions at the
participant’s facility, not Grid Conditions. However, Business Rule 3 does not provide for
recovery of CMSC payments when a load is deviating during a ramp (see Appendix H and the
further discussion in Section 9 below).
Finding #11 (Failure to Ramp):
The occasions during the Relevant Period where the Thunder Bay Facility
failed to ramp after bidding to do so were infrequent. While the resulting
CMSC payments were self-induced and should have been clawed back, the
available evidence does not indicate that the failures to ramp were intentional
attempts by Bowater to exploit a market defect.
7.4.7 Dispatch Deviation in Non-Ramping Hours
During the course of the Investigation, instances of constrained-off CMSC payments arising
from dispatch deviation when the Thunder Bay Facility was not ramping during the Relevant
Period were also noted. When dispatch deviation occurs outside of ramping hours, the IESO
typically claws back the CMSC payments that are triggered pursuant to Business Rule 3.
However, due to a flaw in Business Rule 3, there are occasions where such self-induced CMSC
payments are not clawed back.
Figure 7-6 illustrates an occasion on July 10, 2010 where the Thunder Bay Facility deviated from
its dispatch instruction in a non-ramping hour (HE 23) and the resulting CMSC payments were
146
Responses to RFI, B.11.
PUBLIC VERSION
Products Inc.
100
not fully clawed back. The Thunder Bay Facility’s consumption deviated substantially during
intervals 5 through 12 of HE 23. In all but the last three of these intervals, CMSC payments were
clawed back under Business Rule 3. However, CMSC payments totalling $3,335 were made for
intervals 10, 11 and 12.
During Non-Ramp Dispatch Deviation at the Thunder Bay Facility
July 10, 2010
(MW and $)
The CMSC payments made to Bowater during dispatch deviations in non-ramping hours
occurred infrequently and most of it was clawed back. In certain circumstances, the Facility may
have been returning to its dispatch level after experiencing a consumption decrease, and may
have received CMSC payments that should have been clawed back but was missed by the
dispatch deviation Business Rule. The Panel did not identify evidence indicating an awareness
of, or deliberate attempts to exploit, this particular defect in the Business Rule 3 clawback
formula. Nevertheless, such CMSC payments were unwarranted and should have been clawed
back because they were caused by conditions at the participant’s facility, not Grid Conditions
(see further discussion in Section 9 below).
PUBLIC VERSION
Products Inc.
101
Finding #12 (Constrained-off Dispatch Deviations in NonRamping Hours):
Instances of dispatch deviation in non-ramping hours by the Thunder Bay
Facility during the Relevant Period were infrequent. While the resulting
available evidence does not indicate that these deviations were intentional
attempts by Bowater to exploit a market defect.
7.5
Profits or Benefits to the Market Participant
The mere receipt of a CMSC payment does not necessarily mean that a market participant has
profited or benefited. A market participant will profit or benefit when the CMSC payments that it
receives exceed the reduction in operating profits caused by adhering to an IESO dispatch
instruction to consume less or more electricity than its quantity in the market schedule.
The Panel has analyzed detailed hourly and interval-by-interval data relating to prices,
differences in schedules, the reasons for those differences, and the amount of CMSC payments,
in order to assess whether Bowater profited from the CMSC payments it received. As indicated
in Section 7.4.2 above, the bid prices used by Bowater substantially exceeded the operating
profit reductions resulting when expected consumption was constrained off during ramp down or
ramp up hours on weekdays, and even more so on weekends. Similarly, Bowater self-induced
quantity differences between the market and dispatch schedules by submitting bid quantities in
excess of expected consumption, by its chosen ramp down timing pattern, and by ramping faster
than its submitted ramp rates (see Sections 7.4.3, 7.4.4 and 7.4.5). The evidence summarized in
Section 7.4 clearly shows that Bowater viewed its behaviour and the associated CMSC payments
as sources of incremental profits rather than as compensation for reduced operating profits
caused by responding to dispatches arising from Grid Conditions. With the exception of
occasional failures to ramp and dispatch deviations in non-ramping hours, almost all of
PUBLIC VERSION
Products Inc.
102
Bowater’s CMSC payments was the result of exploiting market defects (Finding #1) by using
high bid prices (Finding #4)147, submitting maximum bid quantities in excess of consumption
(Finding #8)148, expanding schedule quantity differences through ramp down timing (Finding
#9)149 and ramping faster than submitted ramp rates (Finding #10)150.The Panel has determined
that Bowater profited by $11.0 million from the CMSC payments it received.
The CMSC payments to Bowater were earned almost exclusively during ramp down and ramp
up hours. As detailed in Section 7.2.3, during these hours Bowater was receiving CMSC
payments to implement its own self-induced changes in consumption. In fact, the CMSC
payments received were substantially greater than the cost of energy consumed during ramping
hours. During the Relevant Period, Bowater received self-induced CMSC payments in 735 hours
by ramping. During these hours Bowater received $12.0 million in net CMSC (of the total of
$12.3 million in Table 4-2) but paid only $3.2 million in energy charges (including the
applicable Global Adjustment and Uplift amounts). Bowater has not identified any costs or other
reductions in its operating profits arising from responding to IESO dispatches during its
self-induced ramps. It is therefore clear that Bowater profited substantially from the
constrained-off CMSC payments generated by self-induced ramping of the Thunder Bay Facility.
Finding #13 (Profit or Benefit to Bowater):
Bowater profited $11.0 million from the CMSC payments received as a result
of the behaviours set out in Findings #4 and #8 – 10, which exploited the
market defects set out in Finding #1.
147
The Panel estimates a CMSC impact of $10.3 million.
The Panel estimates an incremental CMSC impact of $42,000.
149
The Panel estimates an incremental CMSC impact of $0.6 million.
150
The incremental CMSC impact is subsumed in Finding #4.
148
PUBLIC VERSION
Products Inc.
103
7.6
Expense or Disadvantage to the Market
Net CMSC payments (after the IESO’s clawback procedures are applied) are charged to all
Ontario wholesale electricity market customers as part of Uplift charges. Wholesale market
participants that are distributors ultimately pass these costs on to their customers. When one
participant exploits market defects in the CMSC system and profits from its behaviour, this
imposes an expense and disadvantage throughout the market. All customers bear the cost by
paying higher Uplift charges than would otherwise have been incurred. Indeed Bowater
personnel explicitly recognized that the CMSC payments it was receiving “are shared by all
consumers through uplift charges”.151
Between February and August 2010, Bowater received $12.3 million in net CMSC payments.
The Panel has determined that Bowater profited by $11.0 million from the CMSC payments it
received and increased Uplift charges by $0.12/MWh.152
Finding #14 (Expense or Disadvantage to the Market):
All customers in the wholesale energy market were disadvantaged by paying
additional Uplift charges of $0.12/MWh as a result of Bowater’s behaviours.
7.7
Conclusion
Bowater is a large and sophisticated market participant. The exploitative behaviours identified
above were engaged in with the knowledge of many personnel, including senior management at
Bowater and its ultimate parent company, Abitibi Bowater Inc. Bowater repeatedly and
deliberately engaged in multiple behaviours to exploit market defects in a manner which
triggered substantial CMSC payments for Bowater at the expense of wholesale loads who pay
the Uplift charges in the Ontario wholesale electricity market.
151
Attachment to an email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], April 20, 2010.
152
Total Market Demand between February to August 2010 was approximately 90 TWh.
PUBLIC VERSION
Products Inc.
104
The Panel concludes that during the Relevant Period, four of the behaviours giving rise to $11.0
million in CMSC payments to Bowater were intentional and that the behaviours constituted
gaming: using bid prices well above the Marginal Benefit of Consumption; submitting bid
quantities in excess of consumption levels; the ramp down timing pattern; and ramping faster
than submitted ramp rates.
The Panel did not find the infrequent occasions where Bowater failed to ramp or deviated from
dispatch in non-ramping hours and received constrained-off CMSC payments to be exploitative.
Nevertheless, such CMSC payments were unwarranted and should have been clawed back. In
Section 9 the Panel examines the need for further improvements in the Market Rules and IESO
procedures to prevent unwarranted CMSC payments in the future.
Finding #15 (Finding of Gaming):
Bowater exploited market defects.
In so doing, Bowater received $11.0
million in CMSC payments during the Relevant Period, and there was a
corresponding disadvantage or expense to the market. Bowater’s conduct
constitutes gaming.
PUBLIC VERSION
Products Inc.
105
8.
ABITIBI’S CONDUCT IN RESPECT OF THE FORT FRANCES FACILITY
This section contains the Panel’s assessment of whether Abitibi engaged in gaming in relation to
CMSC payments in respect of the Fort Frances Facility. The introductory sections (Sections 8.1
and 8.2) describe the constrained-off and constrained-on CMSC payments received and Abitibi’s
typical operating pattern for the Facility, including its bidding strategy and ramping pattern. The
subsequent sections (Sections 8.3 to 8.7) assess the four elements of the gaming framework set
out in Section 5.5, namely whether there were market defects which were exploited by Abitibi to
its profit or benefit and to the expense or disadvantage of the market. The Panel concludes
(Section 8.8) that five of Abitibi’s behaviours constituted gaming.
8.1
CMSC Payments to Abitibi
Between January and August 2010, Abitibi received approximately $9.7 million in net CMSC
payments. The gross constrained-off CMSC payments were approximately $18.5 million, of
which $10.7 million was clawed back by the IESO. The gross constrained-on CMSC payments
were approximately $3.7 million, of which approximately $1.8 million was later voluntarily
repaid by Abitibi. Table 8-1 summarizes the CMSC payments, clawbacks and repayment
applicable to the Relevant Period.
Table 8-1: Gross and Net CMSC Payments to Abitibi for the Fort Frances Facility
($000)
Month
January
February
March
April
May
June
July
August
Voluntary
Repayment
Total
Gross ConstrainedOn CMSC
-23
-5
819
606
1,253
656
386
Gross ConstrainedOff CMSC
861
2,481
3,322
1,446
1,588
2,639
2,976
3,231
Clawback of
Constrained-Off CMSC
453
2,068
2,183
408
914
1,203
1,988
1,500
Net
CMSC
385
413
1,134
1,857
1,280
2,689
1,644
2,117
-1,825
-
-
-1,825
$1,867
$18,544
$10,717
$9,694
PUBLIC VERSION
Products Inc.
106
8.2
Typical Operating Pattern
The Fort Frances Facility’s energy consumption pattern in 2010 varied on a daily basis. The
quantity consumed varied, as did the hours, duration and magnitude of ramps. Abitibi explained
its varied operating pattern as the result of ongoing internal changes in the process of making
paper which influenced the amount of load required at different times of the day throughout the
year.153 Abitibi further explained that the permanent idling of one paper machine in April 2009
created excess pulp capacity and a “need to schedule several outages for inventory control during
a 24-hour period of operation.”154
On January 22, 2010, Abitibi registered with the IESO to operate as a “net” load or generator
under the Market Rules.155 Prior to this date, Abitibi was party to a Power Purchase Agreement
(PPA) with the Ontario Electricity Financial Corporation. On January 21, 2010 the PPA expired
and Abitibi was free to combine the load and generator. This change had the effect of reducing
hourly Uplift and Global Adjustment charges, as those charges were then levied on net rather
than gross consumption.156
8.2.1 Operating as a Net Load or Generator
To operate on an aggregated basis, Abitibi submitted offers and bids based on its expected
combined operations. For example, if the generator was expecting to produce ● MW and the
load was expecting to consume ● MW, Abitibi could submit economic offers for ● MW of
generation and an uneconomic bid (or no bid) for the load. Conversely, if the generator was
expecting to produce ● MW and the load was expecting to consume ● MW, Abitibi could submit
an uneconomic offer (or no offer) for the generator and economic bids for ● MW of load. The
153
155
IESO, Market Rules, Chapter 7, Section 2.3 (Issue 21, December 9, 2009).
156
154
PUBLIC VERSION
Products Inc.
107
IESO tools continued to issue separate dispatches for each resource — a zero dispatch for one
and the net portion for the other.157
8.2.2 Bidding Strategy
Abitibi had historically utilized an extremely high bid price for the load at the Fort Frances
Facility. Since 2004, it had regularly bid $●/MWh for approximately ● MW of its load capacity,
with the remaining consumption bid at $2,000/MWh (making the load non-dispatchable). On
January 29, 2010, it changed its regular bid for its net load amount to $●/MWh at all times for
nearly all of its bid capacity, with only ● MW of capacity bid at the non-dispatchable price of
$2,000/MWh.
For select hours between April 7, 2010 and September 2, 2010, Abitibi adopted an extreme
negative bid price strategy by bidding a substantial portion of its net load capacity at -$●/MWh.
A negative bid price suggested that Abitibi was only willing to consume if it was paid $●/MWh
to do so. At times the Facility would bid with $●/MWh and -$●/MWh laminations. The
negative price lamination was bid at a quantity that was no more than 15 MW above the quantity
that was bid at $●/MWh. The extreme low bid price had no effect on the amount of CMSC
payments to Abitibi during ramping (see Section 8.4). However, it is relevant to the
constrained-on CMSC payments received by Abitibi (see Section 8.5).
8.2.3 Ramping Pattern
Abitibi stated that the permanent idling of paper machine number 6 required it to ramp
frequently throughout the day and, in particular, to ramp down when it experienced high pulp
levels but had no capacity to store pulp.158 On both weekdays and weekends, the Fort Frances
Facility would typically ramp multiple times. To implement a ramp down, Abitibi changed its
quantity bid for the net load from a higher level to a lower level. The bid price remained at
157
Settlement statements are still issued for both the generator and the load, but are based on the net output. For
those intervals where the aggregated facility operates as a net load, the load is charged for the net consumption. In
intervals where the facility operates as a net generator, the generator is paid for the net output.
158
PUBLIC VERSION
Products Inc.
108
$●/MWh or $●/MWh in the ramp down hour. To implement a ramp up, Abitibi changed its
quantity bid from a lower level to a higher level. The bid price remained at $●/MWh or
$●/MWh in the ramp up hour.
Prior to combining the load and generator, Abitibi used ramp rates which ramped the
dispatchable load at the Fort Frances Facility up or down in three stages, as summarized in
Table 8-2. When Abitibi aggregated its load and generator, the net load began using modified
ramp rates of ● MW/min at every stage.
Table 8-2: Ramp Rates for the Dispatchable Load at the Fort Frances Facility
January 2009 – August 2010
(MW and MW/min)
RAMP DOWN
RAMP UP
MW Range
Pre-February
2010
February-August
2010
MW
Range
Pre-February
2010
February-August
2010
●159 to ●
● MW/min
● MW/min
● to ●
● MW/min
● MW/min
● to ●
● MW/min
● MW/min
● to ●
● MW/min
● MW/min
● to ●
● MW/min
● MW/min
● to ●
● MW/min
● MW/min
Abitibi explained the change to its ramp rates as follows:
The first lamination of load is rated at ● MW/Min while the
remaining load maintains the slower rate of ● MW/Min. It is
because the load is now netted with the generation at the delivery
point that the first lamination is not seen by the IESO because of
the ● MWh of generation subtracted from the ● MWh of gross
load. The net effect of this is measured at the point of connection,
equal to approximately ● to ● MWh at the applicable ramp rate of
● MW/Min.160
In other words, when bidding as a net load, Abitibi was indicating that all the (net) MWs that
were available for dispatch had a ramp rate of ● MW/min. It did not explain why the
159
160
Abitibi submitted ramp rates for up to ● MW, although it never bid to consume more than ● MW.
PUBLIC VERSION
Products Inc.
109
● MW/Min and ● MW/Min rates were lowered. The relationship between Abitibi’s ramp rates
and the CMSC payments it received is discussed in Sections 8.4.3, 8.4.4, 8.4.5.2 and 8.5.4.
8.2.4 Constrained-off CMSC Payments During Ramping
As discussed in Section 6.3.4 and Appendix G, a dispatchable load can initiate ramping through
a self-induced dispatch by changing the prices and/or quantities that it bids. The same applies
for a net load. Both the market and the constrained schedules will change to allow the facility to
ramp to its desired new level of net consumption. The changes to the load’s dispatch are caused
by the ramping decision manifested in the participant’s bid, not by Grid Conditions.
Nevertheless, CMSC will be paid if the schedule quantities diverge as a result of a change in the
load’s price/quantity bids. The resulting payments are self-induced CMSC.
Like Bowater, Abitibi induced changes in its desired energy consumption by changing its
quantity bid from one hour to the next hour. This in turn triggered self-induced constrained-off
CMSC payments because the quantity in the constrained schedule falls below the market
schedule quantity during the ramp period. Although the frequency and magnitude of ramping
varied from day to day for the Fort Frances Facility, the bid price and ramp rates were generally
consistent. An example of a day where the Fort Frances Facility ramped frequently is Sunday,
March 21, 2010, when Abitibi received over $93,000 in constrained-off CMSC payments (net of
clawbacks). Abitibi bid the net load at $●/MWh in every ramping hour with the ramp rates of ●
MW/min. The ramping of consumption, and the CMSC payments triggered during the ramping
periods, are shown in Figure 8-1.
PUBLIC VERSION
Products Inc.
110
Figure 8-1: Sample Ramping Pattern and CMSC Payments
for the Net Load at the Fort Frances Facility
March 21, 2010
(MW and $)
8.2.4.1 CMSC on Ramp Down
On March 21, 2010 the Fort Frances Facility bid to ramp six times, triggering CMSC payments
each time. Each ramp down and each ramp up took six intervals to complete, resulting in five
intervals where the constrained schedule differed from the unconstrained schedule and the
facility received constrained-off CMSC payments. For example, Abitibi changed its quantity bid
for the net load from ● MW in HE 9 to ● MW in HE 10, indicating that it wanted to ramp down.
The bid price remained at $●/MWh. To accommodate the change in its consumption bid, the
IESO began to dispatch the Fort Frances Facility down beginning in interval 8 of HE 9 using the
maximum submitted ramp down rate of ● MW/minute. To calculate CMSC payments, the IESO
settlement tool took the difference between (i) the market schedule and (ii) the constrained
PUBLIC VERSION
Products Inc.
111
schedules for each interval during the ramp period, and multiplied each quantity difference by
(iii) the $●/MWh bid price less (iv) the MCP for the interval (which was in the $25-30/MWh
range). In the result, nearly $10,000 in CMSC payments were triggered by this particular ramp
down event.
Table 8-3 shows the total energy charges and CMSC payments to Abitibi for each interval in HE
9 on March 21, 2010 when the Fort Frances Facility was ramping down. The net CMSC
payments ($9,744) were substantially larger than the energy charges ($1,244), such that Abitibi
was actually receiving $8,500 while consuming the amount of energy it wanted to during its
self-induced ramp down.
Table 8-3: Energy Charges and CMSC Payments Received on a Typical Ramp Down
of the Net Load at the Fort Frances Facility
March 21, 2010, HE 9
(MW, $/MWh and $)
Interval
Unconstrained
Schedule
(MW)
Constrained
Schedule
(MW)
Actual
Consumption
(MW)
Energy
Charges*
($)
●
●
●
●
1
●
●
●
●
2
●
●
●
●
3
●
●
●
●
4
●
●
●
●
5
●
●
●
●
6
●
●
●
●
7
●
●
●
●
8
●
●
●
●
9
●
●
●
●
10
●
●
●
●
11
●
●
●
●
12
Total
$1,244
** None of the Business Rules applied to claw back CMSC on the ramp down.
MCP
($/MWh)
27.95
25.20
25.91
26.27
26.98
27.43
27.44
27.45
29.45
29.45
29.89
30.49
Bid
Price
($/MWh)
Net CMSC
**
($)
●
●
●
●
●
●
●
●
●
●
●
●
657
1,313
1,868
2,625
3,281
$9,744
8.2.4.2 CMSC on Ramp Up
For HE 11, Abitibi changed its quantity bid for the net load from ● MW to ● MW, indicating
that it wanted to ramp up back to the consumption level it had been at in HE 8. To accommodate
PUBLIC VERSION
Products Inc.
112
Abitibi’s change in consumption bid, the constrained schedule began to dispatch the Fort Frances
Facility up beginning in interval 1 of HE 11 using the submitted ramp up rate of ● MW/minute.
Accordingly, it was dispatched from ● MW to ● MW over six intervals. The market schedule
moved to ● MW in interval one, using the 3x ramp rate multiplier, before moving to ● MW for
the remaining intervals. To calculate CMSC payments, the IESO settlement tool took the
difference between (i) the market schedule and (ii) the constrained schedule for each interval
during the ramp period, and multiplied each quantity difference by (iii) the $●/MWh bid price
less (iv) the MCP for the interval (which was approximately $26/MWh). In the result, over
$7,000 in CMSC payments were triggered by this particular ramp up event.
Table 8-4 shows the total energy charges and CMSC payments to Abitibi for every interval in
HE 11 on March 21, 2010 when the Fort Frances Facility was ramping up. The net CMSC
payments ($7,002) were substantially larger than the energy charges ($1,104), such that Abitibi
was actually receiving nearly $5,900 while consuming the amount of energy it wanted to during
its self-induced ramp up.
PUBLIC VERSION
Products Inc.
113
Table 8-4: Energy Charges and CMSC Payments Received on a Typical Ramp Up
of the Net Load at the Fort Frances Facility
March 21, 2010, HE 11
(MW, $/MWh and $)
Interval
Unconstrained
Schedule
(MW)
Constrained
Schedule
(MW)
Actual
Consumption
(MW)
Energy
Charges*
($)
MCP
($/MWh)
25.55
26.62
26.62
26.62
26.98
26.62
26.98
27.69
26.98
27.34
27.34
25.91
Bid
Price
($/MWh)
Net CMSC
***
($)
1,316
2,630
2,285
●
●
●
●
●
1
●
●
●
●
●
2
●
●
●
●
●
3
●
●
●
●
●
4
●
●
●
●
●
5
●
●
●
●
●
6
●
●
●
●
●
7
●
●
●
●
●
772
8
●
●
●
●
●
9
●
●
●
●
●
10
●
●
●
●
●
11
●
●
●
●
●
12
Total
$1,104
$7,002
*Includes Global Adjustment and Uplift charges.
**CMSC payments of $1,627, $1,348, $740, $575 and $115 in intervals 4, 5, 6, 7 and 9, respectively, were clawed
back under Business Rule 3.
8.2.5 Constrained-on CMSC Payments with Negative Bid Prices
On numerous occasions beginning on April 7, 2010 and continuing to the end of the Relevant
Period, Abitibi submitted a bid price of -$●/MWh for as little as ● MW or as much as ● MW of
consumption by the Fort Frances Facility. The negative bids were most frequently submitted
between HE 7 and HE 21. While a bid at a negative price indicates a strong desire not to
consume (i.e. the load is only willing to consume if paid the amount of the negative bid), the Fort
Frances Facility in fact was often consuming during such hours.
There were two different scenarios in which the Fort Frances Facility consumed energy during
hours for which it submitted negative-price bids:
Constrained-on Consumption (Scenario #1): When the Nodal Price at the Facility fell
below the submitted bid price of -$●/MWh, the (net) load was constrained on.
PUBLIC VERSION
Products Inc.
114
Consumption Deviation (Scenario #2): When the Nodal Price at the Facility was above
the submitted bid price of -$●/MWh and the Facility was scheduled to not be consuming,
it deviated from its dispatch schedule with the result that it appeared to be constrained on.
As noted above, consuming in such circumstances is inconsistent with a highly negative
bid price, which indicates a desire not to consume.
Of the $3.7 million in constrained-on CMSC payments received by Abitibi between April and
August 2010, approximately $1.8 million arose when the Fort Frances Facility was constrained
on while using a negative bid price (scenario #1) and approximately $1.9 million resulted from a
negative bid price combined with deviation from the dispatch schedule (scenario #2).161
A representative example involving both scenarios (during different hours) occurred on June 1,
2010. Figure 8-2 shows the unconstrained and constrained schedules as well as the actual
consumption and CMSC payments for each hour of the day.
161
Abitibi voluntarily repaid $1.825 million in CMSC payments that resulted under scenario #2.
PUBLIC VERSION
Products Inc.
115
Figure 8-2: Schedules and CMSC Payments During Hours with Negative Bid Prices
for the Net Load at the Fort Frances Facility
June 1, 2010
(MW and $)
On this particular day, Abitibi received over $330,000 in constrained-on CMSC payments by
using a bid price of -$●/MWh for the hours of HE 7 through HE 18 inclusive. The CMSC
payments earned in each of the two scenarios identified above is described in the following
sections.
8.2.5.1 CMSC Payments for Constrained-on Consumption (Scenario #1)
In scenario #1, Abitibi received CMSC payments when it was constrained on while using
the -$●/MWh bid price (i.e. when the bid price of -$●/MWh was greater than the Nodal Price for
the Fort Frances Facility but less than the MCP). As shown in Figure 8-2, this scenario resulted
in CMSC payments between the hours of HE 15 and HE 18 inclusive. The constrained-on CMSC
PUBLIC VERSION
Products Inc.
116
payment was equal to MCP + $● (refer to the constrained-on CMSC formula in Section 6.3.2).
As a result, Abitibi received over $170,000 in CMSC payments during this four-hour period.
8.2.5.2 CMSC Payments for “Constrained-on” Consumption Deviation (Scenario #2)
For June 1, 2010 Abitibi submitted net load bids for the Fort Frances Facility as listed in
Table 8-5 below. Beginning in HE 7, it lowered the second lamination quantity and added a
third lamination at -$●/MWh. These bids indicate that (i) Abitibi was treating its first ● MW of
load as non-dispatchable, (ii) it was only willing to be dispatched below ● MW if the Nodal
Price or the MCP rose above $●/MWh, and (iii) it was willing to consume between ● and ● MW
only if it was paid $●/MWh to do so.
Table 8-5: Consumption Bids for the Net Load at the Fort Frances Facility
June 1, 2010
(MW and $/MWh)
Lamination
HE 1 to HE 6
Quantity
1
2
3
HE 7 to HE 18
Price
Quantity
●
●
●
●
-
-
●
●
●
HE 19 to HE 21
Price
Quantity
●
●
●
●
●
●
HE 22 to HE 24
Price
Quantity
●
●
●
Price
●
●
●
●
-
-
Just prior to the beginning of HE 7, the Fort Frances Facility was consuming around ● MW and
was receiving dispatch instructions to reduce consumption toward the ● MW level that Abitibi
had submitted for HE 7. However, the Facility only reduced its consumption to ● MW in HE 7
and HE 8. With the ● MW lamination still in place for HE 9 through HE 18, it actually
increased its consumption to around the ● MW range for the next 10 hours.
The constrained schedule is determined, in part, by the actual consumption of the load. For
every five-minute interval, the IESO dispatch tool re-calculated the Dispatch Envelope within
which it could move the facility, given its submitted ramp rates and starting consumption level.
By constantly consuming well above its dispatch schedule, Abitibi raised the dispatch schedule
upwards (above and away from the market schedule) and the IESO dispatch tool was unable to
dispatch it down to the ● MW quantity bid. However, the market schedule remained at ● MW in
PUBLIC VERSION
Products Inc.
117
response to Abitibi’s submitted bid quantity. In other words, Abitibi induced a divergence
between the dispatch and market schedules and made it appear that the load was constrained on.
Abitibi received self-induced CMSC payments equal to the MCP less -$● (i.e. MCP + $●) for
each MW of difference between the constrained and unconstrained schedules during the hours of
HE 7 and HE 14 inclusive (refer to the constrained-on CMSC formula in Section 6.3.2). As a
result, Abitibi received over $160,000 in CMSC payments by deviating from its dispatch
instructions while using this bid strategy on June 1, 2010.
8.2.6 Representative Pattern of Operation
While the Fort Frances Facility did not have a standard operating pattern repeated each day, the
bid, ramp, consumption and CMSC payments on March 21, 2010 and June 1, 2010 are
illustrative of the manner in which CMSC payments arose at the Fort Frances Facility on
weekdays and weekends during the Relevant Period. The amount of the CMSC payments on any
particular weekday or weekend varied primarily in response to variations in the magnitude and
frequency of ramping, actual consumption relative to scheduled consumption (deviation), and the
MCPs during the applicable ramp and deviation intervals. A summary of Abitibi’s five largest
CMSC payment days at the Fort Frances Facility during the Relevant Period is provided in
Appendix M.
8.3
Defects in Market Rules or Procedures
The market defects discussed in Section 7.3 are also relevant to the constrained-off CMSC
payments received by Abitibi in respect of the Fort Frances Facility. The Panel therefore
reiterates its prior finding, which is applicable to Abitibi as well as Bowater:
PUBLIC VERSION
Products Inc.
118
Finding #1 (Market Defects Related to Constrained-Off
CMSC):
Relevant Period allowed a dispatchable load to receive constrained-off CMSC
payments that exceeded the amount required to compensate for reductions in
operating profits arising from responses to dispatch instructions caused by
Grid Conditions.
An additional defect that is relevant to the CMSC payments made to Abitibi is the lack of IESO
procedures for recovering (or “clawing back”) self-induced constrained-on CMSC payments.
Such payments may provide a dispatchable load with compensation that exceeds the operating
profit reductions arising from being dispatched to consume more electricity than it wanted to.
Moreover, dispatchable loads may be able to self-induce such payments in situations where the
constrained-on dispatch instruction is not the result of Grid Conditions. Although it is not
particularly common for a dispatchable load to be constrained on, the absence of any clawback
rules presented a gap within the Market Rules and the IESO’s CMSC procedures that could be
exploited.
Finding #16 (Market Defects Related to Constrained-on
CMSC):
Relevant Period allowed a dispatchable load to receive constrained-on CMSC
payments that exceeded the amount required to compensate for reductions in
operating profits arising from responses to dispatch instructions caused by
Grid Conditions.
8.4
Exploitation of Constrained-Off CMSC
As set out in Section 5.5, an essential element of gaming is that the market participant engages in
activity which exploits a market defect. The Panel considers that exploitation may exist where
PUBLIC VERSION
Products Inc.
119
the market participant had some level of intention, knowledge or awareness of an opportunity
arising from the market defect. In order to determine whether Abitibi exploited defects in the
CMSC regime, the Panel examined the development of its ramping strategy and the following
specific behaviours, each of which contributed to the large constrained-off CMSC payments
received during ramp periods:
(i)
Abitibi submitted an extremely high bid price for ramping hours, which
increased the amount of the CMSC payment for any difference between
the unconstrained and constrained schedule quantities (see Section 8.4.2).
(ii)
On ramp downs, Abitibi often ramped down faster than its submitted ramp
rates, which reduced its actual consumption and increased the quantity
differences used to calculate CMSC payments (see Section 8.4.3).
(iii)
Abitibi ramped the Fort Frances Facility up and down frequently, which
increased the number of CMSC payments that were received (Section
8.4.4).
(iv)
Abitibi used its generator to respond to self-induced changes in its net
load, which triggered CMSC payments for the net load based on submitted
bid prices and ramp rates that were not reflective of the generator’s
marginal costs and ramping capabilities (see Section 8.4.5).
(v)
Abitibi occasionally failed to ramp up or down in accordance with its bid
and dispatch instructions, which increased the quantity differences that
give rise to CMSC payments (see Section 8.4.6).
(vi)
Abitibi periodically deviated significantly from its dispatch instructions in
non-ramp hours, which triggered CMSC payments that were not always
clawed back under the IESO Business Rules (see Section 8.4.7).
PUBLIC VERSION
Products Inc.
120
8.4.1 Development of the Ramping CMSC Strategy
As early as 2007, Abitibi had actively engaged in developing strategies to increase CMSC
payments during the ramping of the Fort Frances Facility. For example, [Senior Abitibi
Personnel #2], described the implementation of an auto-load shedding program in the
groundwood mill for the purpose of “optimizing on the CMSC revenue”.162 At that time the
IESO was proceeding with a revision of the ramp rate multiplier from 12x to 3x, a change that
would reduce the divergence between the constrained and unconstrained schedules of
dispatchable facilities. [Senior Abitibi Personnel #2] was attuned to the CMSC implications of
such a change and began “the process of determining a new operating strategy to again optimize
on the revenue.”163 Similarly, correspondence between [Senior Abitibi Personnel #2] and
operating personnel contained a detailed analysis with supporting spreadsheet calculations which
showed, for each interval, the megawatts for the constrained and unconstrained schedules, actual
metered energy consumed, and the amount of the CMSC payment that would be received for a
particular ramping strategy. 164
In 2009, Abitibi continued to devise ramping strategies that would trigger CMSC payments as it
made changes to its operations. For example, a number of factors, in particular the permanent
idling of one paper machine (PM6)165, were identified as providing opportunities to engage in
ramping. In an email with the subject “CMSC training”, [Senior Abitibi Personnel #5] referred
to CMSC payments being “scheduled” through ramping:
162
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #2], September 7, 2007. Responses to
RFI, B.16.102.
163
Email from [Senior Abitibi Personnel #2] to [Senior Bowater Personnel #2], September 7, 2007. Responses to
RFI, B.16.102.
164
Emails between [Senior Abitibi Personnel #2] and [AbitibiBowater Inc Personnel #5], March 14, 2007.
165
PUBLIC VERSION
Products Inc.
121
From:
To:
[AbitibiBowater Inc Personnel #2], [AbitibiBowater
Inc Personnel #3], [AbitibiBowater Inc Personnel
#1]
Cc:
Date:
April 17, 2009 03:28 PM
Subject:
CMSC training
With PM6 down we will have more inventory than usual, therefore
we will be able to use this opportunity to do clean up or
maintenance in the department. I would like to train you gents how
to schedule CMSC’s (ramping). If we take the floor down to do
maintenance or cleaning we should always try to “ramp” it down
and up again.166
Similarly, an internal request regarding what “the financial difference to our mill would be if we
were to only have 12 grinders vs 14 grinders available for ramping in 2010?”167 resulted in the
following response:
From:
To:
Cc:
[Senior Abitibi Personnel #6], [Senior Abitibi
Personnel #5]
Date:
December 17, 2009 03:16 PM
Subject:
Re: Groundwood ramping
[Senior Bowater Personnel #6],
A successful ramp of ● mw (● stones) = $20,000 in cmsc payment
If you are not running #5 grinder line, this would reduce the ramp
to ● mw (● stones) = $17,000 in cmsc payment
166
Email from [Senior Bowater Personnel #6] to [Senior Abitibi Personnel #2], December 17, 2009. Responses to
RFI, B.9.4.
167
PUBLIC VERSION
Products Inc.
122
Loss of cmsc per ramp = $3,000
In August we ramped 60 times
But, you should also consider the loss of storage capacity and lost
opportunities to ramp because of this. Maybe [Senior Abitibi
Personnel #5] has better tools to calculate this.168
In 2010 Abitibi continued to analyze CMSC payments arising from plant operating decisions as a
source of profit on a per ramp, per machine basis:
When tank levels allow, we have been manually ramping ● mw of
grwd [groundwood], approx $5500 a ramp…. With #6 grinder line
back on we should be able to ramp ● mw of load, approx $7500 a
ramp… The generator continues to be dispatched on in the
morning and at night. I am going to run an analysis to compare
with the start up and shutdown of the thunder bay ● mw load…
Today the drawbacks are starting to come through, so I will see
how much money is left on the table in all this.169
Personnel at the Fort Frances Facility were also aware that they needed:
… to be careful when scheduling the ramps, from a compliance
perspective. If there are not legitimate reasons to schedule an
outage, this is considered gaming by IESO. They have the right to
remove us from the market and have us pay back the CMSC we
generated. All ramps needs to be justified and should not be
scheduled at the same time every day.170
As discussed in Section 7.4, Abitibi personnel (particularly [Senior Abitibi Personnel #2]) were
also instrumental in advising Bowater’s Thunder Bay Facility on its ramping strategy. This
advice drew upon extensive experience and in-depth knowledge of the relationship between
ramping and CMSC at the Fort Frances Facility. It is clear from the foregoing examples and
other documents provided during the Investigation that Abitibi actively engaged in strategies to
increase its CMSC payments during ramping.
168
Email from [Senior Abitibi Personnel #2] to [Senior Abitibi Personnel #3], February 12, 2010. Responses to RFI,
B.5.1.
170
Email from [Senior Abitibi Personnel #2] to [Senior Abitibi Personnel #1], April 23, 2009. Responses to RFI,
B.9.2.
169
PUBLIC VERSION
Products Inc.
123
Finding #17 (CMSC Ramping Strategy):
Abitibi developed strategies to self-induce CMSC payments at the Fort
Frances Facility, and these were known to senior management.
The main behaviours which triggered large constrained-off CMSC payments for Abitibi are
analysed in the sections below.
8.4.2 Expanding the Magnitude of CMSC Using a High Bid Price
This section examines Abitibi’s knowledge of the operating profit principles underlying the
CMSC regime (Section 8.4.2.1) and the relationship between its Marginal Benefit of
Consumption and bid prices (Sections 8.4.2.2 to 8.4.2.4). The Panel has also analysed the three
explanations for high bid prices that were provided by Abitibi:171
·
Bidding at a very high price reduced the risk of the facility being dispatched
down (which can occur if the Nodal Price is above the bid price) (see Section
8.4.2.5).
·
Bidding at a very high price reduced the risk of being activated to provide OR
(while still being able to obtain revenue from participating in the OR market)
(see Section 8.4.2.6).
§
The Fort Frances Facility as well as the dispatchable load owned by AbitibiConsolidated at Fort William had bid at a similarly high price for a number of
years (see Section 8.4.2.7).
8.4.2.1 Abitibi Understood the Operating Profit Principles in the CMSC Regime
Abitibi and affiliated company personnel understood that the CMSC regime was designed to
compensate for reductions in operating profits based on an assumption that a dispatchable load’s
171
Responses to RFI, B.3, p.2 and B.13 pages 1 and 2.
PUBLIC VERSION
Products Inc.
124
bids would reflect the Marginal Benefit of Consumption. As outlined in Section 7.4.2.1,
personnel at the Fort Frances Facility and the Thunder Bay Facility communicated regularly on
the development of CMSC ramping strategies, including bid prices. [Senior Abitibi Personnel
#2] acted as an advisor to Bowater on an ongoing basis. [Senior Abitibi Personnel #2] was
involved in communications with personnel at the Thunder Bay Facility on the relationship
between bidding strategies and CMSC, the ability to anticipate dispatch instructions, 172 whether
opportunity costs were covered in IESO training materials and other matters.173 Moreover, a
presentation to executives at Abitibi’s parent company, ABI, expressly stated that “[t]he market
rules assume that participants place bids and offers based on their marginal cost and benefit.”174
It is clear that Abitibi understood that the CMSC regime assumed bids by dispatchable loads
would reflect their Marginal Benefit of Consumption.
Finding #18 (Knowledge of CMSC Compensation Principles):
Abitibi was aware that the CMSC regime assumed that dispatchable loads
would bid based on their Marginal Benefit of Consumption and that CMSC
payments were designed to compensate a dispatchable load for operating
profit reductions when it was directed by the IESO to follow a dispatch
different from its market schedule.
8.4.2.2 Bid Prices Exceeded Marginal Benefit of Consumption
In 2009, the Fort Frances Facility permanently idled one of its paper machines (PM6), after
which the Facility was no longer pulp-limited in terms of groundwood. Abitibi stated that the
energy efficiency of the groundwood equipment was maximized by running as fast as possible,
filling up the storage tank and taking short outages. Having to take additional downtime in the
172
RFI, B.2.16.
173
Email from [Senior Bowater Personnel #3] to [Senior Abitibi Personnel #2], [Senior Bowater Personnel #5] and
[Senior Abitibi Personnel #4], June 28, 2010. Responses to RFI, B.13.32. (The IESO training materials in relation to
OR activation are discussed in Sections 7.4.2.7 and 8.4.2.5.)
174
Responses to RFI, B.3.6. See the second slide reproduced at Appendix J.
PUBLIC VERSION
Products Inc.
125
groundwood mill will result in a loss of paper production and an increase in manufacturing
costs.175 In its Responses to RFI, Abitibi calculated that the financial impact of the Fort Frances
Facility being constrained off was $●/MWh in 2010 based on one hour of lost paper production.
However, unlike Bowater, it did not lower its bid prices to this level after the June 2010
communications with the MAU.
Abitibi also indicated that its calculations included fixed costs.176 Fixed costs normally do not
change in response to short term transitory changes in electricity consumption. Including fixed
costs in the Marginal Benefit of Consumption could result in cases where a positive operating
profit that could have contributed to offsetting fixed costs is foregone. Accordingly, they should
not be included in calculating the Marginal Benefit of Consumption. Thus the Abitibi
calculations referenced above overstate the actual Marginal Benefit of Consumption. Since, as
in the case with Bowater, the amount of the overstatement is not readily determinable, the Panel
has used Abitibi’s own calculations as a conservative basis for analyzing whether Abitibi was
bidding at prices which exceeded its Marginal Benefit of Consumption.
In summary, based on Abitibi’s own calculation, the financial impact of one hour of lost
production was no more than $●/MWh in 2010. Thus its bid prices for the net load of $●/MWh
or $●/MWh were well above the Marginal Benefit of Consumption.
8.4.2.3 Marginal Benefit of Consumption is Lower in Self-Induced Ramping Hours
As discussed in Section 7.4.2.2 with respect to Bowater, a Marginal Benefit of Consumption
calculated on a full hour of lost production overstates the financial impact of being dispatched
down during any ramp hour where the facility is only exposed to a fraction of an hour of lost
pulp and paper production.
Abitibi did not demonstrate that the effect of being constrained off while ramping would be
permanently lost mill production, rather than a deferral of production that could be made up later
in the day or week. Given the variations in energy consumption on a daily basis, the Fort
175
176
Responses to RFI, B.8. p.1.
PUBLIC VERSION
Products Inc.
126
Frances Facility generally does not appear to have been operating at capacity on a continuing
basis. Thus any reduced consumption that would arise if constrained off during ramping could
potentially have been recouped soon afterwards.
Even if there would have been lost production as a result of being constrained off, a calculation
based on an hour of lost production overstates the actual financial impact because the Fort
Frances Facility does not operate at full production while ramping up or down. For example,
during ramp downs, when Abitibi self-induced a ramp of ● MW with a ramp rate of ● MW/min,
the Fort Frances Facility was already expecting to reduce consumption of electricity and mill
production in the last six intervals of the hour. Similarly, the Facility was not expecting to
operate at full electricity consumption or mill production during the first six intervals of a ramp
up hour.
Unlike Bowater, Abitibi did not provide data on the portion of ramping hours in which mill
production is occurring. However, the ratio between electricity consumption during a ramping
hour and a full production hour is likely a good and slightly conservative proxy for the level of
mill production during a ramping hour.177 As noted in Section 8.2, the Fort Frances Facility’s
energy consumption pattern in 2010 varied on a daily basis and therefore the ratio between
electricity consumption during a ramping hour and a full hour of production would vary with
each ramp. To estimate this ratio, the Panel considered a commonly used ramp profile by Abitibi,
ramping between ● MW to ● MW of net load at the ramp rates in Table 8-2. Abitibi used this
ramp profile on 376 of its 986 self-induced ramps. During both a ramp up and a ramp down,
electricity consumption during the ramp hour is 80% of consumption during a full hour of
consumption at ●MW. Accordingly, even if there would be permanently lost production as the
result of being constrained off for an entire ramp hour, based on Abitibi’s own estimates above
the operating profit reduction would be no more than $●/MWh ($●*80%).
177
As noted in Section 7.4.2.4, electricity consumption in a normal ramp down hour at the Thunder Bay Facility was
73% of a regular operating hour, which is slightly higher than Bowater’s data that in 65% of a ramp down hour TMP
pulp is still being produced. Similarly, the electricity consumption in a normal ramp up hour was 76% of a regular
operating hour, which is higher than Bowater’s data that in 75% of a ramp up hour TMP pulp is still being produced.
PUBLIC VERSION
Products Inc.
127
8.4.2.4 Abitibi’s Bid Prices Increased CMSC Payments
An approximate estimate of the amount by which Abitibi’s bid prices generated CMSC
payments in excess of operating profit reductions is set out in Table 8-6. This estimate is based
on the conservative assumptions that: (i) there were permanent losses of mill production
resulting from being constrained off (as opposed to deferred production that could be recouped
on subsequent hours or days when the facility was not operating at capacity); (ii) the Marginal
Benefit of Consumption is based on the conservative (maximum) estimates outlined above,
which ignore potential overstatements related to fixed costs; and (iii) all the quantity differences
between the constrained and the greater of the unconstrained schedule and actual consumption
reflected Abitibi responding to IESO dispatch instructions caused by Grid Conditions (which, as
noted in Sections 8.4.3 – 8.4.7, was not, in fact, the case). The total CMSC impact of Abitibi’s
high bid prices is estimated at $5.9 million.
PUBLIC VERSION
Products Inc.
128
Table 8-6: Estimated Impact of Abitibi’s High Bid Prices on
CMSC Payments for the Fort Frances Facility
($/MWh, MWh and $000)
Ramp Up and Down
Bid Price ($/MWh)
●
●
Estimated Marginal Benefit of
Consumption ($/MWh)
●
●
Difference ($/MWh)
1,418
1,369
Constrained-off Quantity
(MWh)*
3,924
229
Total CMSC Impact ($000)
5,564
314
*For intervals where Abitibi received a net CMSC payment.
PUBLIC VERSION
Products Inc.
129
Finding #19 (Operating Profit Impact of Being Constrained
Off):
(a) During periods when Abitibi was not operating the Fort Frances
Facility at capacity, there would be virtually no reduction in operating
profits as a result of being constrained off during a ramping hour because
production could be made up in a subsequent hour.
(b) Even in situations where the Fort Frances Facility was capacity
constrained, Abitibi’s bid prices during the Relevant Period substantially
exceeded its Marginal Benefit of Consumption and the reduction in
operating profits that would result from the net load at the Fort Frances
Facility being constrained off during ramping hours.
(c) Based on data provided by Abitibi and the Facility’s electricity
consumption pattern, the difference between Abitibi’s bid price of $●/MWh
(or $●/MWh) and its Marginal Benefit of Consumption when ramping (up
or down) was at least $1,418/MWh (or $1,369/MWh).
(d) Abitibi’s high bid prices were used to obtain CMSC payments that more
than compensated Abitibi for operating profit reductions by at least $5.9
million.
8.4.2.5 The Risk of Being Constrained Off Did Not Justify Abitibi’s Bid Prices
The Panel has analyzed Bowater’s claim that a high bid price was necessary to prevent being
constrained off during ramping in Section 7.4.2.6. The analysis in respect of Abitibi is similar.
For the reasons discussed in Section 7.4.2.6, in a situation where Abitibi was constrained off
during a ramp, it would receive a CMSC payment. This would offset any negative effect on
PUBLIC VERSION
Products Inc.
130
Abitibi’s operating profits (unless it was using a bid price lower than its Marginal Benefit of
Consumption, which was not the case at any time during the Relevant Period).178
As also noted in Section 7.4.2.6, if Abitibi was genuinely concerned about the risk of being
dispatched down, it could have bid $2,000/MWh to render the facility non-dispatchable during a
ramp down or ramp up hour. By bidding an extra $●/MWh (or $●/MWh, when using a $●/MWh
bid), it could readily have eliminated such a risk but would have not been eligible for CMSC
payments. The fact that it chose not to do so is further confirmation that Abitibi’s assertion that
its very high bid prices were necessary to avoid the risk of being constrained off is not credible.
Abitibi’s assertion that there was a material risk of being constrained off while ramping is also
not credible. The Panel conducted an analysis of Nodal Prices at the Fort Frances Facility during
the Relevant Period and during the immediately preceding year (i.e. between January and
December 2009). The analysis considered Abitibi’s actual bid prices of $●/MWh or $●/MWh, as
well as an alternative bid price that reflected Abitibi’s calculation of a $●/MWh Marginal
Benefit of Consumption (see Section 8.4.2.2) and a further alternative bid price that reflected the
Panel’s estimate of Abitibi’s Marginal Benefit of Consumption during a ramping hour
($●/MWh) (see Section 8.4.2.3). During the 20-month period in question, there were 22,161
five-minute intervals during self-induced ramp hours (10,329 intervals in 2009 and 11,832 in
2010). The results are shown in Table 8-7.
178
In fact, for the reasons discussed in Section 8.4.2.3 the $●/MWh amount exceeded Abitibi’s Marginal Benefit of
Consumption during ramping hours.
PUBLIC VERSION
Products Inc.
131
Table 8-7: Likelihood of the Fort Frances Facility Being Constrained-Off
at Various Bid Prices During Self-Induced Ramping Hours
January 1, 2009 to August 28, 2010
($/MWh, number and % of intervals)
Period
January to
December
2009
January to
August 2010
Bid
Price
($/MWh)
●
●
●
●
●
●
●
●
Number of Intervals Economically Constrained Off
Ramp Down
17
30
36
46
Ramp Down
9
14
14
28
Ramp Up
2
11
12
13
Ramp Up
2
6
6
16
Percent of
Intervals
Economically
Constrained Off
Total
19
41
48
59
Total
11
20
20
44
0.184%
0.397%
0.465%
0.571%
0.093%
0.169%
0.169%
0.372%
This analysis confirms that a high bid price of $●/MWh or $●/MWh was almost never necessary
to prevent the Fort Frances Facility from being dispatched down when ramping. Based on the
outcomes in 2009, which would have been the most recent information available to Abitibi, the
probability of being constrained off was remote. Moreover, the results throughout the Relevant
Period confirm that the probability was remote. Even if Abitibi had bid at a significantly lower
price reflecting its own estimated Marginal Benefit of Consumption, the likelihood of being
constrained off during a ramping hour would have been remote (and any negative impact on
Abitibi’s operating profits would have been compensated by a CMSC payment).
PUBLIC VERSION
Products Inc.
132
Finding #20 (Risk of Being Constrained Off):
The risk of being constrained off during self-induced ramping hours did not
justify Abitibi’s use of a bid price of $●/MWh or $●/MWh, or any other level
above the Marginal Benefit of Consumption of the Fort Frances Facility.
8.4.2.6 The Risk of Operating Reserve Activations Did Not Justify Abitibi’s Bid Prices
As a dispatchable load, Abitibi is eligible to participate in the IESO’s OR market. For most of
the Relevant Period (from January to July 2010), Abitibi did not offer OR for its dispatchable net
load. Accordingly, it is not credible for Abitibi to claim that it was necessary to bid at a high
price to avoid OR activation during those months.
Starting in August 2010 Abitibi began offering between ● MW and ● MW of OR from the Fort
Frances Facility. For the same reasons as noted in Section 7.4.2.7, the Panel rejects Abitibi’s
claim that it was necessary to bid at a high price to avoid OR activation:
·
Contrary to Abitibi’s claim, IESO training materials did not require or instruct
dispatchable loads to bid at high prices to avoid OR activation.
·
If a dispatchable load is activated for OR (during a ramping hour or
otherwise), it would be compensated for its reduced energy consumption
based on its bid price. Thus, as long as Abitibi’s bid price was not lower than
its Marginal Benefit of Consumption (which was never the case during the
Relevant Period), there would be no reduction of operating profit as a result of
an OR activation during a self-induced ramping hour.
·
Abitibi could have eliminated any OR activation risk entirely by either: (i) not
bidding into the OR market during self-induced ramping hours, or (ii) using a
$2,000/MWh (non-dispatchable) energy bid price (which would preclude
being activated to provide OR).
PUBLIC VERSION
Products Inc.
133
·
The actual risk of an OR activation in the Northwest region during a selfinduced ramping hour was remote.
To confirm the level of activation risk, the Panel examined the frequency of all OR activations in
the Northwest region during the hours in which the Fort Frances Facility was ramping up or
down. There were no OR activations in the Northwest region during the Fort Frances Facility’s
self-induced ramping hours in August 2010 (or indeed in any self-induced ramping hour during
Relevant Period). The frequency of all OR activations during the hours in which the Fort Frances
Facility was ramping up or down in 2009 was also examined. There were no activations in the
Northwest during the self-induced ramping hours of the Fort Frances Facility in 2009.179 As
noted above, an internal document prepared for ABI management referred to the historical
experience at the Fort Frances Facility as “twice in 4 years”.180 This confirms that Abitibi knew
that the OR activation risk at the Fort Frances Facility was minimal, and its claim to the contrary
is not credible.
Finding #21 (Risk of Being Activated for Operating Reserve):
The risk of being activated to provide operating reserve during self-induced
ramping hours did not justify Abitibi’s use of an energy market bid price of
$●/MWh or $●/MWh, or any other level above the Marginal Benefit of
Consumption of the Fort Frances Facility.
8.4.2.7 Historical Use of High Bid Prices by Abitibi or Affiliates Did Not Justify Abitibi’s Bid
Prices
Abitibi’s argument that $●/MWh or higher bid prices have been its past practice and has also
been used by another dispatchable load is not a justification for possible gaming behaviour for at
179
The Fort Frances Facility was activated for OR on 3 occasions during non-ramping hours in 2009 and none
during the Relevant Period.
180
See “Thunder Bay 2010 Power Cost – October 1st, 2009” (reproduced in Appendix J). Responses to RFI, B.3.6,
p. 3.
PUBLIC VERSION
Products Inc.
134
least three reasons. First, the fact that Abitibi or other participants engaged in conduct in the past
that may or may not have exploited a market defect does not have any bearing on whether
Abitibi was exploiting a market defect during the Relevant Period. Abitibi reviewed the MSP’s
Monitoring Report on the IESO-Administered Electricity Markets for the period from November
2009 – April 2010 that reported on high CMSC payments to dispatchable loads. A
communication to ABI noted: “It is interesting in the write up below that there is very little
mention of ramping and bidding at Fort Frances. I believe this is because the process at FF is 5-6
years old, well established and condoned by the IESO.”181 The Panel does not agree that the lack
of IESO action or a prior investigation by the MSP in any way condones behaviour that may
constitute gaming. Second, it is possible that other market participants could have a Marginal
Benefit of Consumption equal to or greater than this level, whereas Abitibi does not (see
Sections 8.4.2.2 and 8.4.2.3). Third, it is possible that high bid prices may be used by
dispatchable loads that ramp quickly and therefore trigger negligible CMSC payments.
Finding #22 (High Bid Prices by Other Loads):
The historical use of high bid prices by Abitibi or any other dispatchable
loads does not provide a justification for Abitibi’s high bid prices during selfinduced ramping hours.
8.4.3 Ramping Faster than Submitted Ramp Rates
As discussed in Section 7.4.5, a pattern of ramping faster than submitted ramp rates has the
effect of creating greater divergence between the quantities (constrained schedule or actual
consumption, and the unconstrained schedule) used to calculate CMSC payments. It also
indicates that the submitted ramp rates understate the actual ramping capability of the facility,
which results in a longer ramping period and higher CMSC payments. As noted in Section 7.4.5,
[Senior Abitibi Personnel #2] was well aware that ramping faster than submitted ramp rates
181
Email from [Senior Bowater Personnel #3] to [AbitibiBowater Inc Executive #3], August 31, 2010. Responses to
RFI, B.13.107.
PUBLIC VERSION
Products Inc.
135
increased in CMSC payments, and [Senior Abitibi Personnel #2] advised Bowater on these
issues.182
Abitibi stated that it was forced to ramp down faster than its submitted ramp rates when the Fort
Frances Facility was exposed to “uncontrolled upsets” or “environmental and safety” issues.183
However, the Panel’s analysis indicates that the Fort Frances Facility ramped down faster than
its submitted ramp rates in one or more intervals in 68% (356 of 524) of its ramp downs during
the Relevant Period. This indicates that the net load was being ramped faster than its submitted
ramp rates more frequently than just in response to unexpected operating conditions.
An example of one of the 356 ramp downs where the Fort Frances Facility ramped faster than
submitted ramp rates is March 5, 2010 in HE 3. Abitibi bid $●/MWh in the hour and submitted
the ramp rates in Table 8-2. The ramping of consumption, and the CMSC payments triggered
during the ramping intervals, are shown in Table 8-8. The CMSC payments obtained during the
ramp down ($9,078) exceeded the CMSC payments that would have been received had the Fort
Frances Facility ramped from one dispatch instruction to the next at its submitted ramp rate
($8,241). The Facility ramped faster than its submitted ramp rate in interval 9, moving from ●
MW to ● MW in a five-minute period. According to Abitibi’s submitted ramp rate, the Facility
was only capable of reaching ● MW, and not ● MW, from a starting point of ● MW.184 The
actual ramp rates in intervals 10 and 11 were as submitted, but the faster ramping in interval 9
resulted in further quantity differences between the actual CMSC payment and the payment that
would otherwise have been triggered intervals 10 and 11. Although the MW difference resulting
from faster ramping may appear to be small, each constrained-off MW was being paid the
difference between Abitibi’s bid price of $●/MWh and the MCP (which was less than $25/MWh
during these intervals).
182
See the correspondence between [Senior Abitibi Personnel #2] and [Senior Bowater Personnel #5] between
February 8-16, 2010, reproduced above. Responses to RFI, B.2.14-B.2.16.
183
184
Abitibi’s actual ramp rate in interval 9 of ●MW/minute was 40% faster than its submitted ramp rate of
● MW/minute.
PUBLIC VERSION
Products Inc.
136
Table 8-8: CMSC Payments on a Fast Ramp Down of the Fort Frances Facility
March 5, 2010, HE 3
(MW, $/MWh and $)
Interval
1
2
3
4
5
6
7
8
9
10
11
12
Unconstrained
Schedule
(MW)
Constrained
Schedule
(MW)
Actual
Consumption
(MW)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
MCP
($/MWh)
25.87
Net
CMSC
($)
24.76
13.72
24.51
24.38
24.38
13.72
23.64
386
24.10
1,317
23.76
1,975
13.72
2,492
24.10
2,908
Total
Expected
Consumption
(MW)
Expected
Net CMSC
($)
●
●
●
●
●
●
●
●
●
●
●
●
$9,078
329
987
1,646
2,316
2,962
$8,241
The Panel has not estimated the aggregate incremental impact of Abitibi’s faster ramp down on
CMSC payments. This calculation is partially subsumed in the estimate set out in Section 8.4.2.4
because the estimate in that section is based on the difference between the unconstrained and the
greater of the constrained schedule and the actual quantity consumed, and therefore accounts for
any constrained-off megawatts from fast ramping. The estimate in Section 8.4.2.4, however,
only accounts for the incremental CMSC payments for fast ramping constrained-off megawatts
based on the difference between Abitibi’s actual bid prices and its estimated Marginal Benefit of
Consumption. Fast ramping constrained-off megawatts are entirely self-induced and should not
be compensated for at any bid price. The estimate in Section 8.4.2.4 therefore underestimates the
impact of fast ramping on constrained-off CMSC payments by the number of fast ramping
constrained-off megawatts multiplied by the difference between Abitibi’s estimated Marginal
Benefit of Consumption and the MCP. While the Panel has not undertaken an interval by
interval estimation of the incremental impact of Abitibi’s fast ramping beyond what is already
accounted for in the estimate set out in Section 8.4.2.4, the Panel is nevertheless satisfied that the
PUBLIC VERSION
Products Inc.
137
vast majority of CMSC payments associated with Abitibi’s fast ramping is subsumed in Section
8.4.2.4.
Finding #23 (Ramping Down Faster than Submitted Rates):
a) Abitibi’s Fort Frances Facility was able to, and frequently did, ramp
down the net load faster than its submitted ramp rates, indicating that its
ramp rates were lower than its operational capabilities.
b) The submission of ramp rates that were lower than the Fort Frances
Facility’s operational capabilities increased the magnitude of
constrained-off CMSC payments to Abitibi.
c) The ramping down of the Fort Frances Facility faster than the submitted
ramp rates increased the magnitude of constrained-off CMSC payments
to Abitibi.
8.4.4 Frequent Ramping
During the CCAA restructuring of the Abitibi Bowater entities in 2009, the Fort Frances Facility
shut down paper machine number 6. As a result, the Facility had excess pulp but limited capacity
to store it. Abitibi stated that this required it to ramp frequently throughout the day and, in
particular, to ramp down when it experienced high pulp levels with insufficient further storage
capacity. Documents provided to the Panel during the Investigation indicate that the ramping
strategy was closely connected to the impact on CMSC payments. For example, operating
personnel at the Fort Frances Facility were given “CMSC training” related to the facility’s
inventory conditions and “how to schedule CMSC’s (ramping)”. 185
185
Email from [Senior Abitibi Personnel #5] to [AbitibiBowater Inc Personnel #2], [AbitibiBowater Inc Personnel
#3] and [AbitibiBowater Inc Personnel #1], April 17, 2009. Responses to RFI, B.9.1.
PUBLIC VERSION
Products Inc.
138
A detailed analysis of ramping options and CMSC implications occurred in preparation for
combining the load and generator in late 2009:
From:
To:
Cc:
Personnel #5]
Date:
December 17, 2009 11:34 AM
Subject:
Groundwood ramping
can you please tell me what the financial difference to our mill
would be if we were to only have 12 grinders vs 14 grinders
available for ramping in 2010?
We have a problem with one of our grinders and do not know
whether it will require a rewind and we will have FF6 down in
2010 so the demand for groundwood will be lower – this
information on the lost financial opportunity by having 1 less
grinder motor to ramp will help us make the right business
decision.186 (emphasis added)
186
PUBLIC VERSION
Products Inc.
139
[Senior Abitibi Personnel #2] responded with the financial implications as follows:
From:
To:
Cc:
Personnel #5]
Date:
December 17, 2009 03:16 PM
Subject:
[Senior Bowater Personnel #6],
A successful ramp of ● mw (● stones) = $20,000 in cmsc payment
If you are not running #5 grinder line, this would reduce the ramp
to ● mw (● stones) = $17,000 in cmsc payment
Loss of cmsc per ramp = $3,000
In August we ramped 60 times
But, you should also consider the loss of storage capacity and lost
opportunities to ramp because of this. Maybe [Senior Abitibi
Personnel #5] has better tools to calculate this.187
187
PUBLIC VERSION
Products Inc.
140
In a subsequent portion of the email chain, [Senior Abitibi Personnel #2] concluded as follows:
From:
To:
Cc:
[Senior Bowater Personnel #6], [Senior Abitibi
Personnel #5]
Date:
December 17, 2009 3:59 PM
Subject:
Once the PPA is terminated (jan 20), we will be combining the
load and generation delivery points in order to net out the
monthly/hourly uplifts. Approx savings of $1.5M/month.
This will require a new operating strategy until the grid valve
replacement is done on cogen in June.
The strategy for the ramps is still being determined at this point,
but it is likely they will be done more frequently and only shedding
● mw instead of ● mw.
By not having #5 grinder line running we would be limiting
ourselves on the number of ramps.
We should meet in the new year to discuss.188 (emphasis added)
It is clear from this and other emails that Abitibi viewed CMSC payments as a financial flow that
could be managed and forecasted. This is inconsistent with the design of the CMSC regime,
which is to provide market participants with compensation for unexpected reductions in
operating profits caused by unpredictable Grid Conditions.
In addition to understanding the impact of frequent ramping on CMSC payments, Abitibi
personnel were aware that the frequent ramping strategy could constitute gaming. For example,
six days after the “CMSC training” email (reproduced in Section 8.4.1) was sent, [Senior Abitibi
188
PUBLIC VERSION
Products Inc.
141
Personnel #2] warned colleagues (including [Senior Abitibi Personnel #3] at the time as well as a
member of Abitibi Bowater Inc.’s management) as follows:
From:
To:
Cc:
Personnel #5], [Senior Abitibi Personnel #6],
Date:
April 23, 2009 03:42 PM
Subject:
Re: Groundwood meeting
Hello [Senior Abitibi Personnel #1],
[Senior Abitibi Personnel #5] just stopped by to review what was
discussed at your mtg this afternoon. (I did not receive an
invitation, otherwise I would have been there)
I just want to make sure everyone is aware that we need to be
careful when scheduling the ramps, from a compliance perspective.
If there are not legitimate reasons to schedule an outage, this is
considered gaming by the IESO. They have the right to remove us
from the market and have us pay back the CMSC we generated.
All ramps [need] to be justified and should not be scheduled at the
same time every day.189
The Panel examined the frequency and magnitude of self-induced ramping of more than ● MW
by Abitibi at the Fort Frances Facility over a four-year period.190 The results are shown in Table
8-9.
189
Self-induced ramps of ●MW or less can be achieved in the unconstrained and constrained schedule in one
interval at a ramp rate of ●MW/min and therefore generate no CMSC.
190
PUBLIC VERSION
Products Inc.
142
Table 8-9: Self-Induced Ramps Greater than ● MW at the Fort Frances Facility
2007 to 2010
(MW)
Self-Induced Ramps During the Year
Self-Induced Ramps January-August
Year
#
MWs
#
MWs
2007
524
8,879
323
5,740
2008
492
6,699
382
5,068
2009
710
18,888
588
15,690
2010
1,191
24,286
1,087
22,327
Between January and August 2010, Abitibi self-induced a ramp (either up or down) at the Fort
Frances Facility in 1,087 hours. This represents an 85% increase in frequency from the previous
year and a 150% increase relative to the 2007-2009 average. Similarly, the total MWs of selfinduced ramping during the first 8 months (January-August) of 2010 was 40% higher than in the
first 8 months of 2009 and 150% higher than the January-August average for 2007-2009. As
indicated in emails reproduced above, Abitibi ramped with the knowledge that, in combination
with its high bid price, each ramp would generate significant CMSC payments.
The impact of frequent ramping on CMSC payments depends upon the quantity of affected MWs
and the bid price. Assuming that the percent difference between the January to August 2010
versus 2009 quantities (i.e. 40%) is used as an estimate of increased ramping frequency, the
CMSC impact was approximately $5.8 million191 at Abitibi’s bid prices and would have been
$1.6192 million if Abitibi had been bidding at the estimated Marginal Benefit of Consumption of
$●/MWh (see Section 8.4.2.3) during ramping hours.
191
Calculated based on the ramping data in Table 8-6 and average MCP between January and August, 2010, as:
3,924/1.4*(●-38) + 229/1.4*(●-38).
192
To avoid overlapping calculations with the estimates in Section 8.4.2.4 and in Finding #19, the estimate of the
incremental CMSC impact from Abitibi’s frequent ramping is calculated based on the estimated Marginal Benefit of
PUBLIC VERSION
Products Inc.
143
Finding #24 (Frequent Ramping):
Abitibi increased its CMSC payments through frequent ramping of the Fort
Frances Facility during the Relevant Period by at least $5.8 million.
8.4.5 Combination of Load and Generator
During the course of the Investigation, it was noted that there were instances during the Relevant
Period where Abitibi used its generator at the Fort Frances Facility to respond to a self-induced
ramp in its net load consumption. As a net load, Abitibi can respond to a dispatch instruction to
increase (decrease) consumption either by the load consuming more (less), or by the generator
producing less (more) output. (The effect at the revenue meter of the aggregated facility is the
same.) Where the generator is used for a self-induced ramp, the load does not move, but Abitibi
receives CMSC payments based on the schedule quantity differences for the net load during the
ramping period.
8.4.5.1 Operating Profit Impact of Net Consumption Changes
Constrained-off CMSC payments were made for each constrained-off MW at an amount equal to
the interval MCP less the net load’s bid price of $●/MWh. For example, Figure 8-3 shows the
meter readings of the actual load and generator as well as the actual consumption and the
constrained and unconstrained schedules for the Fort Frances Facility on a net basis on July 28,
2010. From HE 6 to HE 7 Abitibi bid to reduce the consumption of its net load from ● MW to
● MW. To meet the dispatch instruction, it increased the output from its generator rather than
reducing consumption from its load. Similarly, from HE 21 to HE 22 Abitibi bid to increase its
net load from ● MW back up to ● MW and achieved this change by decreasing output from its
generator.
Consumption. If Abitibi had bid its estimated Marginal Benefit of Consumption, the extra CMSC payments related
to the ramp down pattern would have amounted to approximately $1.7 million.
PUBLIC VERSION
Products Inc.
144
Figure 8-3: Schedules and CMSC Payments when Using the Generator
to Ramp Changes in the Net Load at the Fort Frances Facility
July 28, 2010
(MW and $)
Abitibi received CMSC payments for both these self-induced ramps. The net amounts after
clawback were $1,700 for the morning ramp down and $2,300 for the evening ramp up.
Where the generator effects a change in the net load, the impact on Abitibi’s operating profits is
a function of the marginal cost of the generator rather than the Marginal Benefit of Consumption
(since the electricity consumed by, and presumably the pulp and paper produced by, the Fort
Frances Facility are not affected). In its Responses to RFIs, Abitibi stated that the marginal cost
of the generator was $●/MWh. 193 Using this amount in the CMSC calculation for the morning
193
Responses to RFI, B.7.48. In 2009, prior to combining the generator with the load, Abitibi generally submitted a
bid price of -$●/MWh for the output of the generator at the Fort Frances Facility, indicating its intention to generate
in almost all market conditions.
PUBLIC VERSION
Products Inc.
145
ramp down would have resulted in CMSC payments of $42 (a reduction of $1,658, or about
97%). If Abitibi had bid at its estimate of the generator’s marginal cost ($●/MWh) for the
evening ramp up hour, it would have resulted in the net load being uneconomic in the market
schedule in certain intervals (because the MCP ranged between $80/MWh to $90/MWh) or being
constrained off (because the Nodal Price was greater than $86/MWh). The Fort Frances
Facility’s generator would have had to delay its ramp down until HE 2 — which is what would
be expected to occur when a generator is economic based on market conditions.194 The
consumption of electricity to operate the load at the Fort Frances Facility would not have been
affected, and the CMSC payments would have been negligible.
The behaviour referred to above occurred on an infrequent basis. In the absence of evidence
indicating Abitibi deliberately exploited a market defect by submitting the load’s bid of $●/MWh
for a ramp up or down of the generator, the Panel has concluded that this behaviour was not
exploitative, although the CMSC payments that were triggered were unwarranted.
194
If there were reasons why the generator needed to stop operating in HE 21, or had increased costs for operating
in HE 22, those could justify a higher bid price, but it is unlikely that such a price would be anywhere near the bid
prices that Abitibi was using. For a discussion of similar issues related to CMSC payments to generators on selfinduced ramp downs, see Market Surveillance Panel, Monitoring Document on Generator Offer Prices Used to
Signal an Intention to Come Offline, August 19, 2011,
online: http://www.ontarioenergyboard.ca/OEB/Industry/About%20the%20OEB/Electricity%20Market%20Surveill
ance/Monitoring%20Document%20-%20Generator%20Offers.
PUBLIC VERSION
Products Inc.
146
Finding #25 (Bid Prices When Using Generator to Alter Net
Consumption):
When Abitibi used the generator to implement self-induced changes to the net
load at the Fort Frances Facility, the bid prices it submitted did not reflect
the marginal cost of the generating facility, resulting in CMSC payments that
substantially exceeded the amount needed to compensate Abitibi for any
operating profit reductions, but was not a deliberate attempt to exploit a
market defect.
8.4.5.2 Ramp Rates for Changes to the Net Load
Similarly, where the generator is used to implement a self-induced ramp, the generator’s ramp
rates rather than the load’s ramp rates reflect the pace at which the net load would be able to
change. Abitibi consistently submitted a ramp rate of ● MW/min (up and down) for the entire
capacity of the generator during the Relevant Period. If the generator ramp rate had been
submitted by Abitibi for the changes in the net load discussed previously in this section, which
were implemented by adjusting generator output, no CMSC payments would have been made
because the constrained and unconstrained schedules would have achieved the entire ramps in a
single interval. In the absence of evidence indicating Abitibi deliberately exploited a market
defect by submitting ramp rates that did not reflect the actual ramping of the net load, the Panel
has concluded that this behaviour was not exploitative, although the CMSC payments that were
triggered were unwarranted.
PUBLIC VERSION
Products Inc.
147
Finding #26 (Ramp Rates When Using Generator to Alter Net
Consumption):
When Abitibi used the generator to implement self-induced changes to the
net load at the Fort Frances Facility, the ramp rates it submitted did not
reflect the actual ramping capabilities of the generator, resulting in CMSC
payments that substantially exceeded the amount needed to compensate
Abitibi for any operating profit reductions, but was not a deliberate attempt to
exploit a market defect.
8.4.6 Failure to Ramp
During the course of the Investigation, it was noted that there were occasional situations during
the Relevant Period where the Fort Frances Facility failed to ramp up or down, even though
Abitibi had submitted bid quantities that indicated it wanted to increase or decrease energy
consumption. Such failures to ramp resulted in discrepancies between the market and dispatch
schedules, and therefore triggered constrained-off CMSC payments.
For example, on August 16, 2010, in HE 7, Abitibi bid for the net load to ramp up, but then
failed to follow the dispatch schedule that reflected the planned ramp for that hour as well as HE
8 and HE 9. The scheduled and actual consumption, as well as the CMSC payments triggered
during the planned ramping periods, are shown in Figure 8-4.
PUBLIC VERSION
Products Inc.
148
for the Net Load at the Fort Frances Facility During a Failure to Ramp
August 16, 2010
(MW and $)
The market schedule was determined by the load’s submitted bid price and quantity, and moved
up to the load’s identified consumption level of ● MW. However, the dispatch schedule can only
move within the Dispatch Envelope.195 Because the load did not increase its energy
consumption, the dispatch schedule remained at around ● MW. Had the Fort Frances Facility
followed its dispatch instructions: (i) the dispatch schedule would have increased toward the
market schedule; (ii) there would have been smaller quantity differences between the two
schedules; and (iii) smaller CMSC payments would have been triggered. While portions of the
CMSC payments were clawed back, the quantity difference resulting from the failure to ramp
triggered over $37,000 in net CMSC payments during HE 7 and HE 9.
Abitibi stated that the failures to follow dispatch instructions occurred when the Fort Frances
Facility experienced equipment failures.196 This behaviour occurred on an infrequent basis. In
195
196
See the discussion in Section 7.4.6.
PUBLIC VERSION
Products Inc.
149
the absence of evidence indicating that such failures to ramp were deliberate, the Panel has
concluded that this behaviour was not an attempt to exploit a market defect. Even though the
CMSC payments triggered when the net load failed to ramp did not arise from exploitative
conduct, they were unwarranted and should have been clawed back because they were caused by
conditions at the participant’s facility, not Grid Conditions. However, Business Rule 3 does not
provide for the recovery of CMSC payments where a load is deviating during a ramp (see
Appendix H and the further discussion in Section 9 below).
Finding #27 (Failure to Ramp):
The occasions during the Relevant Period where the Fort Frances Facility
failed to ramp after bidding to do so were infrequent. While the resulting
available evidence does not indicate that these failures to ramp were
attempts by Abitibi to exploit a market defect.
8.4.7 Dispatch Deviation in Non-Ramping Hours
During the course of the Investigation, it was noted that there were instances of constrained-off
CMSC payments during non-ramping hours in the Relevant Period where the Fort Frances
Facility deviated from intended consumption and/or the market schedule. As discussed in
Section 7.4.7 above, when dispatch deviation occurs outside of ramping hours the IESO typically
claws back the CMSC payments that are generated, although as a result of a flaw in Business
Rule 3 this does not always happen.
Figure 8-5 illustrates an occasion on March 14, 2010 where the Fort Frances Facility deviated
from its dispatch instructions between HE 7 and HE 18. The vast majority of the CMSC
payments were clawed back. However, in HE 18, CMSC payments totalling $1,721 were made
for intervals 2 and 3, and were not clawed back under Business Rule 3.
PUBLIC VERSION
Products Inc.
150
Figure 8-5: Schedules and CMSC Payments During
Non-Ramp Dispatch Deviation at the Fort Frances Facility
March 14, 2010
(MW and $)
The net CMSC payments to Abitibi during non-ramp dispatch deviations were modest and
haphazard. The Panel did not identify evidence indicating an awareness of, or deliberate attempts
to exploit, this particular defect in the Business Rule 3 clawback formula. Nevertheless, as noted
in Section 7.4.7, such CMSC payments were unwarranted and should have been clawed back
because they were caused by conditions at the participant’s facility, not Grid Conditions (see
further discussion in Section 9 below).
PUBLIC VERSION
Products Inc.
151
Finding #28 (Constrained-Off Dispatch Deviations in NonRamping Hours):
Instances of dispatch deviation in non-ramping hours by the Fort Frances
Facility were infrequent. While the CMSC payments were self-induced and
should have been clawed back, the available evidence does not indicate that
these deviations were intentional attempts by Abitibi to exploit a market
defect.
8.5
Exploitation of Constrained-On CMSC
The large amounts of generation compared to demand within the Northwest region, as well as the
limited transmission connections between this region and the rest of Ontario and neighbouring
jurisdictions, often results in very low or negative Nodal Prices.197 In its January 2010
Monitoring Report, the Panel noted that distorted price signals in the Northwest created potential
opportunities for market participants “to obtain excessive CMSC payments from the marketplace
through strategic bidding practices” and that there was a risk that market participants could
“game the market.”198 Three months after the publication of the Panel’s report, Abitibi began to
receive large constrained-on CMSC payments through the adoption of a negative-price bidding
strategy in certain hours for the Fort Frances Facility. In total it received constrained-on CMSC
payments of $3.7 million over five months, $1.8 million for constrained-on consumption
(Scenario #1 from Section 8.2.5.1) and $1.9 million for consumption deviation (Scenario #2
from Section 8.2.5.2). Abitibi subsequently voluntarily repaid $1.825 million of the latter amount
(see Section 8.5.1).
197
See, e.g., the discussion in Market Surveillance Panel, Monitoring Report on the IESO-Administered Electricity
Markets for the period from May 2009 to Oct 2009, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/msp_report_201001.pdf (the “January 2010 Monitoring
Report”), p. 89.
198
Ibid., p. 101.
PUBLIC VERSION
Products Inc.
152
The Panel examined the development of Abitibi’s negative-price bidding strategy (Section 8.5.2)
and assessed whether the large constrained-on CMSC payments received by Abitibi resulted
from the following behaviours:
·
Bidding an extremely negative price of -$●/MWh, which contributed to
large CMSC payments when the Fort Frances Facility was constrained on
(see section) Section 8.5.3.
·
Deviating from dispatch instructions, which triggered large constrained-on
CMSC payments that were not subject to claw back under the Business
Rules (see Section 8.5.4).
8.5.1 Repayment of Portions of the Constrained-on CMSC
In discussions with the MAU that occurred prior to the initiation of the Investigation regarding
constrained-on CMSC payments at the Fort Frances Facility, Abitibi asserted that its deviations
from dispatch were unintentional, as were the associated CMSC payments. Abitibi therefore
agreed to repay the portions of constrained-on CMSC payments arising from dispatch deviation
during scheduling changes at 6:00 a.m. and 6:00 p.m.:
With respect to the constrained on payments at Fort Frances, it
should be noted that these payments are not a pure windfall
brought about by the reclassification of the net load/gen as stated
in your e-mail, but as a result of a change in the bidding strategy
during on peak hours. During times when the shadow price drops
below the bid price, the load is constrained on to consume while
the generator is constrained off. Fort Frances follows these
dispatches both on the load and generator, resulting in legitimate
constrained on payments for the load. During the June 18 phone
call it was pointed out that the load was not following its
dispatches during schedule changes (6 AM and 6 PM). Fort
Frances has since changed the operating procedure to follow
dispatches during these types of schedule changes eliminating the
deviation from dispatch. Abitibi agreed with the IESO on the call
that a portion of the CMSC generated during the constrained on
PUBLIC VERSION
Products Inc.
153
periods should be clawed back. But this is not to say all of it
($2.7M) should be clawed back. …199 (emphasis added)
Abitibi calculated the constrained-on CMSC associated with these dispatch deviations as
follows:200
Fort Frances CMSC Payments
Constrained ON During Negative Bid Pricing
Legit
Non-Legit
April
$71,996.50
$740,886.64
May
$238,198.39
$384,462.69
June
$611,713.58
$621,469.23
July
$543,068.88
$78,191.47
Total
$1,464,977.35 $1,825,010.03
The $1.825 million amount was repaid by Abitibi through an adjustment to its September 2010
invoice from the IESO. However, the repayment does not preclude the Panel from assessing
whether the behaviour constituted gaming, and that assessment is discussed in Section 8.5.4.
The Panel also notes that Abitibi continued to bid at -$●/MWh throughout the month of August
2010 and into September 2010 and received more than $15,000 in further constrained-on CMSC
payments in analogous situations. Abitibi received nearly $1.924 million in constrained-on
CMSC payments as a result of dispatch deviation. The $100,000 difference between Abitibi’s
repayment and the amount it received has not been repaid.
Abitibi also received CMSC payments when the Nodal Price fell below the net load’s submitted
bid price of -$●/MWh. Abitibi advised the MAU that it intended to retain all CMSC payments
which were triggered when “a portion of the load was bid in at -$●, and the shadow price
triggered the dispatch to the generator to decrease production while the load received the signal
to increase consumption and both resources were able to follow dispatch. This is the CMSC we
199
Excerpt of email from [Senior Abitibi Personnel #2] to MAU, August 17, 2010. Responses to RFI, B.13.27.
Responses to RFI, B.13.85. The CMSC payments that Abitibi considered to be “non-legit”, because they arose
from consumption deviation during the Fort Frances Facility’s schedule changes, relate to “Scenario #2” described
in Section 8.2.5.2.
200
PUBLIC VERSION
Products Inc.
154
believe to be appropriate”.201 Whether this behaviour constituted gaming is assessed in
Section 8.5.3.
8.5.2 Development of the Negative-Price Bidding Strategy
On March 31, 2010, Abitibi personnel attended an IESO Stakeholder Advisory Committee
(SAC) meeting where the Chair of the MSP made a presentation regarding the Panel’s January
2010 Monitoring Report, including the prevalence of negative Nodal Prices in the Northwest and
the Panel’s concerns that these conditions could provide dispatchable resources with an
opportunity to bid strategically in order to obtain significant constrained-on CMSC payments.202
The Panel recommended that the IESO revise the constrained-on CMSC payment calculation
when market participants bid at a negative price:
Recommendation 3-4: The Panel recommends that, for the
purposes of calculating Congestion Management Settlement Credit
(CMSC) payments, the IESO should revise its constrained-on
payment calculation using a replacement bid (such as $0/MWh)
when market participants (both exporters and dispatchable loads)
bid at a negative price. This would create more consistent
treatment with generators and importers that are constrained-off.203
The January 2010 Monitoring Report stated, and the MSP Chair’s presentation reiterated, that
the constrained-on CMSC payments in the Northwest as a result of negative Nodal Prices were
an unintended consequence of the two-schedule design of the market, and created a potential risk
that dispatchable resources could game the market.204
One week after the SAC meeting, Abitibi introduced a new lamination in its bids for the Fort
Frances Facility: for the ● to ● MW quantity range, it submitted a bid price of -$●/MWh. [Senior
201
Email from [Senior Abitibi Personnel #2] to MAU, August 17, 2010. Responses to RFI, B.13.28.
Market Surveillance Panel, Monitoring Report on the IESO-Administered Electricity Markets for the period from
May 2009 to Oct 2009, online: http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/msp_report_201001.pdf,
section 3.1.
203
Ibid., p. 104.
204
Ibid., pp. 100-105.
202
PUBLIC VERSION
Products Inc.
155
Abitibi Personnel #2] subsequently described the bidding strategy to [AbitibiBowater Inc
Executive #1] as follows:
From:
To:
Date:
06/10/2010 09:30 AM
Subject:
Re: Clearing Prices
The strategy:
Bid ● MWh of load during on peak times at -$●/MWh.
Whenever there is too much generation in the area the load will be
dispatched on to consume anywhere from ● to ● MWh of power
from the grid. This generates the large CMSC revenue.
When the load is not being dispatched on to consume, they are still
able to consume because of a 15MWh deadband the IESO has
given them. The result is that they consume power without any
CMSC payments while still respecting all market rules.
After the market rule change (September), all dispatchable loads
bidding negative will be automatically reverted to $0/MWh. After
this change, the load will only receive the Market Clearing price
when dispatched on to consume.205 (emphasis added)
This email was forwarded to other parent company executives with the comment that “If we
don’t want [Senior Abitibi Personnel #2] to pursue this method and would prefer to be more
conservative, we will need to inform [Senior Abitibi Personnel #2] of our decision/concerns.”206
After receiving this email, [AbitibiBowater Inc Executive #3], followed up with [Senior Abitibi
Personnel #4] and [Senior Bowater Personnel #3], asking: “Do you think this is good practice or
205
Email from [AbitibiBowater Inc Executive #1] to [AbitibiBowater Inc Executive #3], [AbitibiBowater Inc
Executive #4], [AbitibiBowater Inc Executive #3] and [AbitibiBowater Inc Executive #2], June 10, 2010. Responses
to RFI, B.13.4.
206
PUBLIC VERSION
Products Inc.
156
is it going too far? I just want to make sure we do not do something that attracts too much
attention and [puts] in jeopardy what we now use (ramps) in this program.”207
Senior management corresponded about the appropriateness of the constrained-on CMSC
strategy by email as follows:
From:
Sent:
06/15/2010 03:30 PM
To:
Cc:
Subject:
Re: Tr: Clearing prices – Ontario
[AbitibiBowater Inc Executive #3], can we have a discussion about
this?
[Senior Abitibi Personnel #2] just copied me on a note from
[AbitibiBowater Inc Executive #1] that says we should stop this
practice.
After spending the morning getting brought up to speed on the way
we are using the program – I disagree.
In April and May we generated a CMSC of $309,000 and so far in
June $480,000. The reason we changed our bidding structure on
April 7th was the result of [Senior Abitibi Personnel #2] having a
discussion with someone from the IESO. The way I personally
look at it is that when we are constrained on, we are forced to buy
and sometimes we are forced to reduce generation from Biomass.
In other words, stop using an asset to its full potential.
The other side of this story is $1,140,000 created in April and May,
and $535,000 so far in June that is self created CMSC that we
should not consider to be ours.
If we are going to create any attention through this process, we
have already done so. Why stop now?208 (emphasis added)
207
Email from [AbitibiBowater Inc Executive #3] to [Senior Bowater Personnel #3] and [Senior Abitibi Personnel
#4], June 14, 2010. Responses to RFI, B.13.4. The negative-price bid strategy was also described to [Senior Abitibi
Personnel #4] in an email from [Senior Abitibi Personnel #2], June 4, 2010. Responses to RFI, B.9.22.
208
Responses to RFI, B. 13.4.
PUBLIC VERSION
Products Inc.
157
[AbitibiBowater Inc Executive #3] responded to the email on June 15, 2010:
From:
To:
c.c.:
Date:
06/15/2010 07:04 PM
Subject:
Re: Tr: Clearing prices – Ontario
No problem. Just want to make sure we are not putting at risk our
ramps for a temporary higher benefit. Call [Senior Bowater
Personnel #3] and schedule a conf call to get an alignment. Invite
[AbitibiBowater Inc Executive #1] and [Senior Abitibi Personnel
#2] if available. I am open tomorrow from 2 to 5 pm or Thurs
am.209
It is clear from these emails that: (i) Abitibi recognized that a market defect in the constrained-on
CMSC regime would allow it to obtain additional CMSC payments; and (ii) Abitibi consciously
exploited this opportunity by “[bidding] ● MWh of load during on peak times at -$●MWh…
This generates the large CMSC revenue.” It is also clear that the strategy was known to and
effectively condoned by senior management at Abitibi and its parent company. Moreover,
Abitibi knew that these CMSC payments were not compensating for reductions in operating
profits resulting from responding to dispatch instructions caused by Grid Conditions.
On June 11, 2010, the MAU reminded Abitibi that gaming can be found by the Panel where
there is “exploitation of opportunities to profit or benefit from defects in the design of the
market, from poorly specified rules or procedures, or from circumstances that are not expressly
covered by Market Rules or procedures…”.210 However, Abitibi maintained its position that the
new bid strategy with the -$●/MWh lamination was justified because it was not contravening the
Market Rules. In its Responses to RFI, Abitibi also stated that:
209
Responses to RFI, B. 13.4.
Email from MAU to [Senior Abitibi Personnel #2], June 11, 2010, citing the MSP’s Monitoring Document:
Monitoring of Offers & Bids in the IESO-Administered Markets, online:
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MSP_Monitoring_Offers_Bids_Document_20100310.p
df. Responses to RFI, B.13.1.
210
PUBLIC VERSION
Products Inc.
158
The standard bidding practice in Fort Frances had been to bid all
load at a high price to avoid being constrained off during peak
production times. However it was learned that in the northwest
region there is potential to bid the load at a low price in times when
there is too much generation in the area which could trigger the
IESO to constrain the load on, thereby creating CMSC payments.
It was at the March 31, 2010 Stakeholders Advisory Committee
(SAC) meeting where [the Chair of the Market Surveillance Panel]
presented the May – October 2009 MSP report which addressed
constrained-on payment in the Northwest area. After the
presentation [Senior Abitibi Personnel #2] had a conversation with
[Senior IESO personnel] at which time [Senior Abitibi Personnel
#2] questioned dispatchable loads bidding negative in the
northwest. The response from [Senior IESO personnel] was
presented back in a question. “Why would you not bid negative?”
[Senior IESO personnel] then proceeded to explain in detail the
mechanics of bidding negative in the market. Shortly following
this discussion, Fort Frances changed their bid structure where a
portion of their load would be bid in at a negative price.211
The Panel asked the MAU to interview the [Senior IESO personnel]. [Senior IESO personnel]
indicated the following:
·
[Senior IESO personnel] did not specifically recall conversations with Abitibi
personnel about this issue at the SAC meeting on March 31, 2010.
·
[Senior IESO personnel] does not recall telling Abitibi personnel how to bid
or offer. As a general rule, [Senior IESO personnel] does not advise market
participants on how to bid or offer. If asked, [Senior IESO personnel]’s
normal practice is to be forthcoming and answer questions about the
implications of certain bidding/offering behaviour and the mechanics of
payments.
211
PUBLIC VERSION
Products Inc.
159
·
[Senior IESO personnel] recalled that Abitibi was very active during the
dispatchable loads rule amendment process in the fall of 2010212, and those
stakeholdering discussions would have included detailed descriptions and
implications of bidding in certain ways.
As noted above, Abitibi’s Responses to RFI stated that, when asked by [Senior Abitibi Personnel
#2] about bidding at a negative price in the Northwest, [Senior IESO personnel] responded
“[w]hy would you not bid negative?” during the SAC meeting – implying that the IESO was
counselling the behaviour or at least endorsing it. However, [Senior Abitibi Personnel #2]’s
notes from that meeting do not record this statement and do not demonstrate encouragement or
approval of a strategy designed to obtain larger CMSC payments. They are not inconsistent with
[Senior IESO personnel]’s statement that he would not go beyond forthright explanations of the
implications of certain bidding/offering behaviour or the mechanics of payments. 213 [Senior
Abitibi Personnel #2]’s internal follow up emails also do not contain references to being
encouraged or authorized by the IESO to adopt the constrained-on CMSC strategy.
Even if the [Senior IESO personnel] had made the comment that [Senior Abitibi Personnel #2]’s
claims was made, it does not constitute a defence or justification for gaming behaviour. Market
participants are ultimately responsible for their actions in or affecting the wholesale market, not
only in respect of compliance with the Market Rules but also in respect of conduct that may be
subject to review by the Panel for gaming or abuse of market power. The [Senior IESO
personnel] does not have authority to determine what conduct would or would not constitute
gaming.
When assessing the exploitation element of gaming, the Panel will consider all relevant evidence
relating to the decision-making by the market participant that engaged in the conduct. As a
general observation, the Panel notes that, if a market participant intends to use IESO advice as a
212
Referring to the IESO’s stakeholder engagement Constrained-off Congestion Management Settlement Credits for
Dispatchable Loads (SE-89).
213
Responses to RFI, B.11.6. For example, [Senior Abitibi Personnel #2] notes indicate that a $●/MWh offer would
reflect a bid to consume, and that the constrained payments would be high, whereas a $2,000/MWh bid would make
the facility non-dispatchable.
PUBLIC VERSION
Products Inc.
160
basis for conduct it has engaged in, it would be prudent to ensure that the precise questions and
IESO responses are comprehensively and contemporaneously documented. In this case, the
overall evidence, including the emails reproduced or summarized above, indicates that Abitibi’s
decision to adopt the new strategy for obtaining constrained-on CMSC payments was not based
on authorization or encouragement from the IESO, nor was it based on any business reasons
relating to the operation of the Fort Frances Facility. The Panel also notes that, based again on
the overall evidence, [Senior Abitibi Personnel #2] repeatedly led other initiatives by both
Abitibi and Bowater to exploit other market defects in order to increase CMSC payments.214
It was clear from the presentation at the SAC meeting that the possibility of self-induced CMSC
payments arising from negative-price bids was an unintended consequence of the two-schedule
design of the market, that the Panel regarded these payments as unwarranted, and that the IESO
was taking steps to eliminate them.215 Nevertheless, Abitibi continued to exploit the situation
before the Market Rules changed:
From:
To:
Date:
June 15, 2010 10:52 AM
Re: TR: FW: High CMSC Payments made to Abitibi Facilities
Fyi.
I still think we need to be careful not to go too far (negative
bidding) with the power programs as we might attract too much
attention and [lose] what we now have. This is especially true as, if
I understood well, it is only a temporary hole in the program that
will disappear soon.216 (emphasis added)
214
See the various correspondence and other documents cited in Sections 7 and 8 of this report.
Market Rule Amendment Submission to the Technical Panel, Limiting Constrained-on CMSC Payments for
Exporters and Dispatchable Loads, online:
http://www.ieso.ca/imoweb/pubs/icms/tp/2010/05/IESOTP_236_4b_MR_00370_Q00_Amendment_Submission.doc
.
216
Responses to RFI, B.13.3. See also the various correspondence reproduced earlier in this Section.
215
PUBLIC VERSION
Products Inc.
161
The Market Rules were subsequently amended by the IESO to provide for use of a replacement
bid of -$50/MWh for the purposes of calculating constrained-on CMSC payments for all
dispatchable load transactions.217 Abitibi stopped using negative-price laminations in its bid
submissions for the Fort Frances Facility on September 2, 2010. Abitibi has not used a negativeprice lamination since that time. This is a further indication that the -$●/MWh bidding strategy
was deliberate behaviour adopted to exploit a market defect.
In its Responses to RFI, Abitibi acknowledges that it understood that the introduction of a
negative-price bid lamination for a load in the northwest could trigger constrained-on CMSC
payments.218 The -$●/MWh lamination did not coincide with any associated cost of being
constrained on to consume additional electricity. To the contrary, as noted above, Abitibi
intended to deviate from its submitted bids by consuming within the 15 MW Compliance
Deadband when it was not constrained on.
Finding # 29 (Constrained-On CMSC Payment Strategy):
Abitibi’s adoption of a negative-price bidding strategy between April and
August 2010, which was known to senior management, was a deliberate
attempt to exploit a market defect in the CMSC regime that had been publicly
identified as such by the Panel and was in the process of being rectified by
the IESO.
8.5.3 Expanding the Magnitude of CMSC Using a Low Bid Price
Dispatchable loads have an opportunity to receive constrained-on CMSC payments that more
than compensate them for operating profit reductions if they bid at prices that understate their
Marginal Benefit of Consumption in circumstances where the Nodal Price falls below the load’s
bid price. For example, assume a load bids to consume 1 MW at -$1,999/MWh and the MCP is
217
Market Rules, Chapter 9, Section 3.6.5A, added by MR-00370, online:
http://www.ieso.ca/Documents/Amend/mr2010/MR-00370-R00-BA.pdf.
218
PUBLIC VERSION
Products Inc.
162
$30/MWh. This bid implies that the load is only willing to consume an additional MW for an
hour if it is paid $1,999 to do so. In other words, there is a cost, not a benefit, associated with the
additional consumption. If the Nodal Price falls below -$1,999/MWh, the load will be
constrained on and will pay the MCP for the energy it uses; but it will also receive a CMSC
payment equal to the difference between the bid price and the MCP multiplied by the MWs that
were constrained on for each interval (i.e. $2,029 for 1 MW for 1 hour).
If the load’s Marginal Benefit of Consumption for additional energy is greater
than -$1,999/MWh, the CMSC payment will more than compensate the load for any reduction in
operating profits caused by being forced to consume. For example, if the load actually receives a
marginal benefit of $200/MWh from consuming an additional MW, but bids -$1,999/MWh and
is constrained on, it will receive a CMSC payment of $2,029 and pay the $30 MCP, even though
it would have been kept whole by paying $200 for its additional consumption. The load actually
receives an increase in its operating profits in the amount of $2,199 by bidding -$1,999/MWh
and being constrained on to consume.
The evidence referenced above indicates that Abitibi was profiting when it switched a portion of
the bid quantity for the Fort Frances Facility to -$●/MWh. It is clear from the description of the
strategy that Abitibi placed a positive value on consumption even when it was
bidding -$●/MWh. In fact, Abitibi planned on consuming within the 15 MW Compliance
Deadband when it was not constrained on. It also claimed in its Responses to RFI that one hour
of lost paper production would reduce its operating profits by $●/MWh.219 Therefore, when the
load bid at -$●/MWh and was constrained on (or appeared to be constrained on due to its own
deviations, as discussed in Section 8.5.4), Abitibi was actually paid to consume electricity it
wanted to consume and would, based on its own calculations, have benefited by consuming at
any price below $●/MWh. Abitibi consumed as it intended to consume, there was no operating
profit reduction resulting from responding to a dispatch instruction caused by Grid Conditions,
and therefore no need for a CMSC payment to compensate Abitibi.
219
Responses to RFI, B.8.1. (As indicated in Section 8.4.2.2 above, the Panel believes that this calculation overstates
the actual Marginal Benefit of Consumption, but that does not negate the fact that Abitibi was submitting a bid price
which was inconsistent with its own claims regarding the benefit of incremental energy consumption.)
PUBLIC VERSION
Products Inc.
163
Finding #30 (Negative Bid Prices):
a) Abitibi’s -$●/MWh bid price was well below its Marginal Benefit of
Consumption during the hours in which such bids were submitted for the
net load at the Fort Frances Facility.
b) Abitibi’s low bid price was used to obtain CMSC payments that more
than compensated Abitibi for operating profit reductions by at least $1.8
million.
8.5.4 Deviating from Dispatch Instructions to Appear to be Constrained On
When a load deviates from its market schedule but remains within its Compliance Deadband, it
can effectively manipulate the dispatch schedule without being exposed to sanctions for
breaching a Market Rule. It can submit a specific bid quantity that will be used to determine the
market schedule, then deviate from the submitted quantity (by consuming more) and thereby
force the dispatch schedule above and away from the market schedule and toward its actual
consumption level. The market schedule is not updated where a load deviates from its submitted
quantity bids, and the dispatch schedule will not move the load toward its submitted quantity bid
where the load consumes at a level that is higher than its Dispatch Envelope.220 This creates an
opportunity for dispatchable loads to trigger self-induced constrained-on CMSC payments that
are not subject to clawback. 221
Between April 7 and August 19, 2010, the Fort Frances Facility often consumed significantly
more energy than the quantity it bid at extremely negative prices, thereby pulling the dispatch
220
The dispatch algorithm determines the dispatch schedule by taking a load’s actual consumption over the last 10
minutes and calculating the range within which it can dispatch the load, based on the load’s submitted ramp rates.
This Dispatch Envelope indicates a load’s physical capacity to move in any given direction within a five-minute
interval, and the dispatch algorithm cannot dispatch a load outside the Dispatch Envelope.
221
The Market Rules provide for recovery of constrained-off CMSC payments where dispatch deviation has
occurred, (Business Rule 3 describes the specific circumstances in which the IESO would recover constrained-off
CMSC payments where there has been dispatch deviation, pursuant to Market Rules Chapter 9, Section 3.5.1A).
However, this was not expected to be needed for constrained-on situations and neither the Market Rules nor the
Business Rules provide for the recovery of constrained-on CMSC payments when a load is deviating from dispatch.
PUBLIC VERSION
Products Inc.
164
schedule above the market schedule and making it appear that the Facility was constrained on to
consume more electricity than its bid quantity. Abitibi received constrained-on CMSC payments
in 2,457 intervals where the Nodal Price was greater than Abitibi’s bid price (indicating Abitibi
was consuming more MWs than the constrained schedule was directing). For example, as shown
in Figure 8-2 Abitibi received dispatch instructions to consume roughly ● MW between HE 7
and HE 14, yet it was consuming in the neighbourhood of ● MW.222 According to Abitibi’s bids,
it was only willing to consume between ● MW and ● MW if it was paid $●/MWh to do so (i.e. a
bid of -$●/MWh). During this period the Nodal Price was in the $20/MWh to $70/MWh range,
indicating it was not economic for the Fort Frances Facility to be consuming at the level that it
was. The constrained-on quantities (i.e. the differences between the constrained on schedule and
the market schedule) in each interval attracted a CMSC payment equal to the difference between
the MCP (which was in the range of $30/MWh to $50/MWh in these hours) and the -$●/MWh
bid. During these eight hours, such CMSC payments totaled approximately $150,000.
Finding # 31 (Constrained-On Dispatch Deviations):
Abitibi deviated from the Fort Frances Facility’s dispatch instructions on
numerous occasions that resulted in the Facility appearing to be constrained
on and receiving CMSC payments when its bids indicated it did not want to
consume.
As noted in Section 8.5.1, Abitibi has admitted that the constrained-on CMSC payments arising
from its consumption deviations while bidding at -$●/MWh were “non-legit” and it made a
voluntary repayment of 95% of the amount in question. There is no justification for Abitibi to
retain the approximately $100,000 of similar CMSC payments, especially those payments that it
received after it had conceded that CMSC payments of this type were non-legitimate.
222
During this period Abitibi had bid ● MW at $2,000/MWh (making this quantity non-dispatchable), ● MW at
$●/MWh and ● MW at -$●/MWh.
PUBLIC VERSION
Products Inc.
165
8.6
Profits or Benefits to the Market Participant
As the Panel observed in Section 7.5, a market participant will profit or benefit from CMSC
payments when those payments exceed the reduction in the participant’s operating profits caused
by its adherence to an IESO dispatch instruction to consume less or more electricity than its
quantity in the market schedule. The Panel has analyzed (separately for constrained-off and
constrained-on CMSC payments) data relating to prices, differences between the unconstrained
and constrained schedules and the reasons for those differences, the operating and bidding
behaviours of Abitibi that resulted, as well as the associated CMSC payments.
8.6.1 Constrained-off CMSC Payments
As indicated in Section 8.4, the Panel has found that Abitibi submitted bid prices in excess of its
Marginal Benefit of Consumption, and has also engaged in frequent ramping and ramping faster
than submitted ramp rates. Each of these behaviours triggered constrained-off CMSC payments
that more than compensated for operating profit reductions. With the exception of occasional
failures to ramp and dispatch deviations in non-ramping hours, almost all of Abitibi’s
constrained-off CMSC payments was the result of exploiting market defects (Findings #1)
through using a high bid price (Finding #19)223, ramping faster than submitted ramp rates
(Finding #23)224 and frequent ramping (Finding #24)225. The Panel has determined that Abitibi
profited by $7.5 million from the CMSC payments it received.
From January 1 to August 28, 2010, Abitibi received approximately $7.8 million in net
constrained-off CMSC payments. The vast majority of these payments were earned during rampdown and ramp-up hours. As was detailed in Section 8.2.4, the CMSC payments received were
substantially greater than the energy charges (including applicable Global Adjustment and Uplift
amounts) during ramping hours. This resulted in Abitibi actually being paid to consume while
implementing its self-induced ramps. For example, during the Relevant Period Abitibi received
223
225
224
PUBLIC VERSION
Products Inc.
166
self-induced constrained-off CMSC payments in 986 hours by changing its maximum quantity
bid. During these hours Abitibi earned $6.7 million in net CMSC payments but paid only $1.2
million for the electricity it consumed (including applicable Global Adjustment and Uplift
charges). Abitibi has not identified any costs or other reductions in operating profits resulting
from its self-induced ramps. It is clear that Abitibi profited significantly from the constrained-off
self-induced ramping of the Fort Frances Facility.
Finding #32 (Profit or Benefit to Abitibi from Constrained-Off
CMSC):
Abitibi profited $7.5 million from the constrained-off CMSC payments
received as a result of the behaviours set out in Findings #19, 23 and 24,
which exploited the market defects set out in Finding #1.
8.6.2 Constrained-on CMSC Payments
In Section 8.5 the Panel has found that, at various times, Abitibi submitted -$●/MWh bid prices
that substantially understated its Marginal Benefit of Consumption and also consumed at levels
which deviated from its dispatch instructions. These behaviours were used to obtain constrainedon CMSC payments that more than compensated for operating profit reductions. Abitibi has
claimed that such deviations were an “oversight”.226 However, the evidence presented in Section
8.5 shows that the deviations were frequent and the strategy was intentional.
From January 1, 2010 to August 28, 2010, Abitibi received approximately $3.7 million in
constrained-on CMSC payments at the Fort Frances Facility. As discussed in Section 8.2.5, the
constrained-on CMSC payments arose in two different situations:
·
During constrained-on consumption (Scenario #1) Abitibi received CMSC
payments when the Nodal Price fell below the net load’s submitted bid price
226
Responses to RFI B.11.5.
PUBLIC VERSION
Products Inc.
167
of -$●/MWh and the load was constrained on. Abitibi received approximately
$1.769 million in self-induced constrained-on CMSC payments in such situations.
·
During consumption deviation (Scenario #2) Abitibi received self-induced CMSC
payments when it deviated from its dispatch and therefore appeared to be
constrained on. Abitibi received approximately $1.923 million in constrained-on
CMSC payments in these situations. Abitibi paid back $1.825 million but retained
approximately $100,000 of net constrained-on CMSC payments related to
consumption deviation. The voluntary repayment after being contacted by the
MAU does not negate the fact that, at the time the behaviours occurred, Abitibi
was profiting from them.
Abitibi’s internal correspondence and actual consumption patterns clearly indicate that it was
intending to consume electricity regardless of whether or not it was constrained on while bidding
at -$●/MWh, and that the highly negative bid price did not reflect the financial impact of Abitibi
being constrained on to consume. The resulting constrained-on CMSC payments did not
compensate Abitibi for reductions in operating profits resulting from responding to dispatch
instructions caused by Grid Conditions; rather, they provided Abitibi with incremental operating
profits. It is therefore clear that Abitibi’s constrained-on CMSC payments were the result of
exploiting market defects (Findings #1 and #16) through using low bid prices (Findings #30 and
#31). 227 The Panel has determined that Abitibi profited by $1.9 million from the constrained-on
CMSC payments generated by the negative-price bidding strategy.
227
PUBLIC VERSION
Products Inc.
168
Finding #33 (Profit or Benefit to Abitibi from Constrained-on
CMSC):
Abitibi profited $1.9 million from the constrained-on CMSC payments
received as a result of the behaviours set out in Findings #30 and 31, which
exploited the market defects set out in Findings #1 and #16.
8.7
Expense or Disadvantage to the Market
The Panel has already noted in Section 7.6 that net CMSC payments (that is, payments after the
IESO’s clawback procedures and voluntary repayments are applied) are charged to all Ontario
wholesale market customers as part of Uplift charges. This is the case for constrained-on as well
as constrained-off CMSC payments. As a result, when one market participant exploits market
defects in the CMSC system and profits from its behaviour, this imposes an expense and
disadvantage throughout the market. All customers bear the cost by paying higher Uplift charges
than would otherwise have been incurred.
Between January and August 2010, Abitibi received $9.7 million in net CMSC payments (net of
clawbacks and voluntary repayments). The Panel estimates that $ 9.4 million of that total was
self-induced and increased Uplift charges by $0.09/MWh.228 With the exception of occasional
failures to ramp and dispatch deviations in non-ramping hours, almost all of the self-induced
CMSC payments were the result of exploiting market defects (Findings #1 and #16) through
using a high bid price (Finding #19)229, ramping faster than submitted ramp rates (Finding
#23)230, frequent ramping (Finding #24)231, and using low bid prices (Findings #30 and #31).232
228
Total Ontario Market Demand from January to August 2010 was 105 TWh.
230
231
232
229
PUBLIC VERSION
Products Inc.
169
Finding #34 (Expense or Disadvantage to the Market):
All customers in the wholesale energy market were disadvantaged by paying
additional Uplift charges of $0.09/MWh as a result of Abitibi’s behaviours.
8.8
Conclusion
Abitibi is a large and sophisticated market participant. The exploitative behaviours identified
above were engaged in with the knowledge of many personnel including senior management at
Abitibi and its ultimate parent company, ABI. Abitibi repeatedly and deliberately engaged in
multiple behaviours to exploit market defects in a manner which triggered substantial CMSC
payments for Abitibi at the expense of wholesale loads who pay Uplift charges in the Ontario
market.
The Panel concludes that, during the Relevant Period, five behaviours giving rise to $9.4 million
in CMSC payments to Abitibi were intentional and that the behaviours constituted gaming: using
bid prices well above the Marginal Benefit of Consumption (or generator marginal costs) during
self-induced ramps; ramping down faster than submitted ramp rates (including in certain
situations using the fast ramping capability of the generator for ramping of the net load when
load ramp rates were submitted); frequent ramping; the use of a -$●/MWh bid price lamination
in certain hours; and deviating from the dispatch schedule so that the facility would appear to be
constrained on. The latter two behaviours are particularly glaring given that the market defect in
question had been publicly identified as such by the Panel and was being addressed by the IESO.
The Panel did not find the infrequent occasions where Abitibi failed to ramp or deviated from
dispatch in non-ramping hours and received constrained-off CMSC payments to be exploitative.
Nevertheless, such CMSC payments were unwarranted and should have been clawed back. In
Section 9 the Panel examines the need for further improvements in the Market Rules and IESO
procedures to prevent unwarranted CMSC payments in the future.
PUBLIC VERSION
Products Inc.
170
Finding #35 (Finding of Gaming):
Abitibi exploited market defects. In so doing, Abitibi received $9.4 million in
CMSC payments during the Relevant Period, and there was a corresponding
disadvantage or expense to the market. Abitibi’s conduct constitutes gaming.
PUBLIC VERSION
Inc.
171
9.
REVIEW
OF
CONTINUING
CMSC
PAYMENTS
AND
RECENT
DEVELOPMENTS REGARDING REMEDIAL ACTION FOR GAMING
9.1
Constrained-Off CMSC Payments
On August 27, 2010, the IESO Board of Directors enacted an Urgent Market Rule Amendment
which suspended all CMSC payments to constrained-off dispatchable loads in light of the fact
that significant CMSC payments had been made to two dispatchable loads which the IESO
believed to be inconsistent with the intent of the CMSC regime.233
On December 3, 2010 the Urgent Market Rule Amendment was rescinded and the Market Rules
were amended to permanently eliminate constrained-off CMSC payments made to dispatchable
load facilities associated with self-induced ramping. In the introduction to the amendment, the
IESO stated:
It is proposed that dispatchable loads will not be entitled to constrained-off CMSC
payments related to ramping, where such payments are caused by conditions
and/or actions at the load facility, and not by conditions on the IESO-controlled
grid.234
To implement this amendment to the Market Rules, the IESO made a change to its settlement
procedures. If a dispatchable load changes either element of its P/Q Pair from one hour to the
next, and this change triggers ramping, any CMSC payments made for the hour are to be clawed
back.235 Accordingly, dispatchable loads no longer receive constrained-off CMSC payments for
self-induced ramping. The amendment does not prevent CMSC payments from being calculated
or being paid on a gross basis. Instead, it introduces a new clawback mechanism so that, net of
automated clawbacks, the dispatchable loads do not receive ramp-related CMSC payments.
233
See MR-00373, online: http://www.ieso.ca/Documents/mr/MR_00373-R00.pdf.
See MR-00374, online: http://www.ieso.ca/Documents/Amend/mr2010/MR-00374-R00-BA.pdf, p. 2.
235
IESO, Market Manual 5, Part 5.5: Physical Markets Settlement Statements, s. 1.6.9.3.
234
PUBLIC VERSION
Products Inc.
172
9.2
Constrained-On CMSC Payments
On November 11, 2010, the IESO implemented an amendment to the Market Rules which
largely eliminated deviation-induced constrained-on CMSC payments.236 The amendment
established a replacement bid price of -$50/MWh to cap the amount of the constrained-on CMSC
payments to dispatchable loads and -$125/MWh for constrained-on CMSC payments to
exporters. The replacement bid price for dispatchable loads was revised upwards to -$15/MWh
in March 2012 because they were no longer being charged the Global Adjustment for each MWh
of consumption.237
This amendment substantially reduced, but has not eliminated, constrained-on CMSC payments
for dispatchable loads. For example, if a dispatchable load bids a quantity of 10 MW
at -$1,999/MWh and it is constrained on in an hour when the MCP is $20/MWh, it will receive a
CMSC payment of $35 ($20 – (-$15) *10MW) (considerably less than the $20,190 ($20 –
(-$1,999) *10MW) than would have been received before the amendment took effect).
9.3
Continuing CMSC Payments
Table 9-1 summarizes the CMSC payments to all dispatchable loads in 2011 to 2013, with 2009
and 2010 provided for comparison. The data indicates that the vast majority of CMSC payments
to Bowater and Abitibi from 2011 to 2013 were clawed back, although they still collectively
received over $1.25 million in 2011, $1.75 million in 2012 and $1.00 million in 2013. In
aggregate there was in excess of $5 million per year in net CMSC payments being made to
dispatchable loads in 2011 and again in 2012, reflecting the fact that only about 61% of the gross
CMSC payments were clawed back. In 2013 there was a significant increase in net CMSC
payments being made to dispatchable loads, in excess of $13 million, and only about 50% of the
gross CMSC payments were clawed back. The sharp rise in net CMSC payments in 2013 is of
some concern to the Panel.
236
For further details, see MR-00370, online: http://www.ieso.ca/Documents/Amend/mr2010/MR-00370-R00BA.pdf.
237
See IESO, Market Manual 5: Settlements. Part 5.5: Physical Markets Settlements Statements, p. vi, online:
http://www.ieso.ca/imoweb/pubs/settlements/se_RTEStatements.pdf.
PUBLIC VERSION
Products Inc.
173
Table 9-1: Gross and Net CMSC for the Thunder Bay Facility,
the Fort Frances Facility and All Other Dispatchable Loads
2008 – 2013
($000)
Year
Bowater
(Thunder Bay Facility)
Gross
CMSC
Clawback
Abitibi
(Fort Frances Facility)
All Other Dispatchable Loads
Net
CMSC
Gross
ClawNet
Gross
ClawNet
CMSC
back
CMSC
CMSC
back
CMSC
9,487
7,587
1,899
8,942
6,415
2,527
2008
Not Dispatchable
20,527 14,220
6,307
7,111
5,282
1,829
2009
22,312
9,946
12,366
27,814 18,097
9,717
11,327
9,953
1,374
2010
11,979 11,033
946
16,603 16,271
332
11,359
6,071
5,288
2011
12,435 11,705
731
15,660 14,611
1,049
15,500 10,365
5,136
2012
9
8
1*
14,559 13,536
1,023**
26,725 13,255
13,299
2013*
* On December 10, 2012 Bowater stopped bidding into the market, with the exception of 18 hours over March 3 and
March 4, 2013, which designates the facility as non-dispatchable and ineligible for CMSC.
**On September 12, 2013 Abitibi stopped bidding into the market, which designates the facility as non-dispatchable
and ineligible for CMSC.
The Panel considers that the continuing magnitude of net CMSC payments under the current
Market Rules and IESO Business Rules, which equated to incremental Uplift charges payable by
all Ontario wholesale market customers of $0.04/MWh in 2011 and 2012, and $0.09/MWh in
2013,238 is significant enough to warrant further analysis. During the stakeholder consultation
with dispatchable loads in October 2010 that reinstated constrained-on CMSC payments, the
IESO proposed to review “the broader issue of CMSC” in the longer term.239 In the Panel’s view,
such a review is appropriate today.
When the IESO first automated the clawback of CMSC payments in 2007 it reported 6% less
clawback by the automated process compared to the former manual process.240 While substantial
improvements have been made, automated clawbacks are inherently limited in their ability to
238
Based on total wholesale demand of 154 TWh in 2011, 156 TWh in 2012 and 155TWh in 2013.
See SE-89 Session Notes from October 14, 2010, available online at:
http://www.ieso.ca/Documents/consult/se89/se89-20101014-notes.pdf.
240
IESO Powerpoint Presentation: Dispatchable Load CMSC Clawback, December 12, 2006, online
http://www.ieso.ca/imoweb/pubs/consult/dlwg/dlwg-20061212-DL_CMSC%20clawback.pdf p.22
239
PUBLIC VERSION
Products Inc.
174
claw back unwarranted CMSC payments because they fail to consider all possible ways in which
CMSC payments can be triggered. It may be the case that unwarranted CMSC payments are still
being made in part due to limitations in the IESO’s automated clawback rules.
The Panel therefore believes that it would be prudent for the IESO to review the remaining
sources of CMSC payments with a view to determining which, if any, are unwarranted and
should therefore be eliminated. 241 It may also be useful to consider, as an alternative to further
clawback refinements, a revised approach in which CMSC payments are made only where they
are demonstrated to be warranted (i.e. linked to Grid Conditions) rather than relying on an afterthe-fact clawback mechanism.
Recommendation (Review of Continuing CMSC Payments):
a) The IESO should review the CMSC payments being made to dispatchable
loads since the November/December 2010 amendments to the Market Rules
in order to determine whether there are significant amounts that continue to
be unwarranted (i.e., paid as a result of market participant actions rather
than to compensate for operating profit reductions arising from responding
to dispatch instructions caused by Grid Conditions).
b) If necessary, the IESO should make further amendments to the Market
Rules to eliminate unwarranted CMSC payments to dispatchable loads.
9.4
Recent Developments Regarding Remedial Action for Gaming
As set out in Section 5.1, the Panel’s responsibilities include monitoring, investigations and
reporting.242 The Panel submits its investigation reports to the OEB and the IESO. The Panel’s
241
The Panel also noted that it would be useful to undertake such a review in its Monitoring Report on the IESOElectricity Markets for the period from May 2011 – October 2011, online :
http://www.ontarioenergyboard.ca/OEB/_Documents/MSP/MSP_Report_20120427.pdf, pp. 47 and 48.
242
See Electricity Act, 1998, section 37 and OEB By-law, Articles 4, 5 and 7.
PUBLIC VERSION
Products Inc.
175
investigation reports may include recommendations, including recommendations regarding
Market Rule amendments. However, the Panel does not have the legislative mandate to impose
sanctions or remedies when it finds that gaming has occurred. While a compliance and
enforcement regime exists in relation to breaches of the Market Rules, gaming does not
necessarily constitute a breach of the Market Rules. At present, there is no provision in the
Market Rules that addresses gaming as a separate and distinct activity, although as noted below a
“general conduct rule” is currently under development by the IESO.
The Panel regards gaming as a serious concern because of the potential negative impact on the
operation of the wholesale market, the harm to market participants (and ultimately to all
electricity consumers in Ontario) who bear the cost of it, and the undermining of public
confidence in the market. The Panel therefore believes that remedial action should be available
in appropriate cases, whether that action be in the form of penalties, the recovery of gains made
by the market participant, or some other sanction. In addition to remedying conduct that has
occurred, the prospect of meaningful remedial action would help to deter gaming and contribute
to the integrity of the electricity market.
The IESO is currently engaging in stakeholder consultations regarding the introduction of a
“general conduct rule” into the Market Rules that could encompass gaming (among other
matters).243 The Panel supports this initiative, and encourages the IESO to proceed
expeditiously with its consultations and to ensure that any rule that it implements captures the
kinds of conduct that are the subject of this Report or that have been discussed in other Panel
reports.
243
Details of the consultation, referred to as Stakeholder Engagement SE-112, are available at
http://www.ieso.ca/Pages/Participate/Stakeholder-Engagement/SE-112.aspx.
PUBLIC VERSION
Products Inc.
176
10.
SUMMARY OF FINDINGS AND RECOMMENDATION
10.1
Findings – Bowater
The Panel’s findings with respect to its investigation of Bowater are as follows:
Table 10-1: Summary of Findings Related to Bowater and the Thunder Bay Facility
Page
No.
Subject
Finding
1
Market
Defects
Related to
Constrainedoff CMSC
The CMSC rules, formulas and clawback procedures that
existed during the Relevant Period allowed a dispatchable load
to receive constrained-off CMSC payments that exceeded the
amount required to compensate for reductions in operating
profits arising from responses to dispatch instructions caused
by Grid Conditions.
52
2
CMSC
Ramping
Strategy
Bowater developed strategies to self-induce CMSC payments
at the Thunder Bay Facility, and these were known to senior
management.
61
3
Knowledge of
CMSC
Compensation
Principles
Bowater was aware that the CMSC regime assumed that
dispatchable loads would bid based on their Marginal Benefit
of Consumption and that CMSC payments were designed to
compensate a dispatchable load for operating profit reductions
when it was directed by the IESO to follow a dispatch different
from its market schedule.
64
PUBLIC VERSION
Products Inc.
177
No.
Subject
4
Operating
Profit Impact
of Being
Constrained
Off
Page
Finding
72
a) During periods when Bowater was not operating the
Thunder Bay Facility at capacity, there would be
virtually no reduction in operating profits as a result of
being constrained off during a ramping hour because
production could be made up in a subsequent hour.
b) Even in situations where the Thunder Bay Facility was
capacity constrained, Bowater’s bid prices between
February and August 2010 substantially exceeded its
Marginal Benefit of Consumption and the reduction in
operating profits that would result from the Thunder
Bay Facility being constrained off during ramping
hours.
c) Based on data provided by Bowater, the difference
between Bowater’s February - June 2010 bid price of
$●/MWh and its Marginal Benefit of Consumption
when ramping down was at least $1,644/MWh on
weekdays and $1,934/MWh on weekends.
d) Based on data provided by Bowater, the difference
between Bowater’s February - June 2010 bid price of
when ramping up was at least $1,589/MWh on
weekdays and $1,924/MWh on weekends.
e) Based on data provided by Bowater, the difference
between Bowater’s July - August 2010 bid price of
when ramping down was at least $445/MWh on
weekdays and $735/MWh on weekends.
f) Based on data provided by Bowater, the difference
between Bowater’s July - August 2010 bid price of
when ramping up was at least $390/MWh on weekdays
and $725/MWh on weekends.
g) Bowater’s high bid prices were used to obtain CMSC
payments that more than compensated Bowater for
operating profit reductions by at least $10.3 million.
PUBLIC VERSION
Products Inc.
178
Page
No.
Subject
Finding
5
Risk of Being
Constrained
Off
The risk of being constrained off during self-induced ramping
hours did not justify Bowater’s use of a bid price of $●/MWh
or $●/MWh, or any other level above the Marginal Benefit of
Consumption of the Thunder Bay Facility.
78
6
Risk of Being
Activated for
Operating
Reserve
The risk of being activated to provide operating reserve during
self-induced ramping hours did not justify Bowater’s use of an
energy market bid price of $●/MWh or $●/MWh, or any other
level above the Marginal Benefit of Consumption of the
Thunder Bay Facility.
81
7
High Bid
Prices by
Other Loads
The historical use of high bid prices by other dispatchable
loads does not provide a justification for Bowater’s high bid
prices during self-induced ramping hours.
81
8
Maximum Bid
Quantity
a) Bowater’s change in its maximum bid quantity from ●
MW to ● MW from February 19 to May 11, 2010 was
undertaken to, and did, increase constrained–off CMSC
payments.
84
b) The estimated amount of incremental CMSC payments
derived from Bowater’s use of a ● MW maximum bid
quantity at the prices Bowater was bidding was
$330,000.
9
Ramp Down
Timing
a) Bowater used a ramp down pattern for its auxiliaries
that triggered increased CMSC payments during the
Relevant Period when there was a known alternative
ramping pattern that would have generated
significantly lower CMSC payments and that was
compatible with the Thunder Bay Facility’s operational
requirements (having been used before and considered
for use after the Relevant Period).
b) The estimated amount of incremental CMSC payments
derived from Bowater’s ramp down pattern was $3.9
million at the prices Bowater was bidding.
PUBLIC VERSION
91
Products Inc.
179
No.
Subject
10
Ramping
Down Faster
than
Submitted
Rates
Page
Finding
a) Bowater’s Thunder Bay Facility was able to, and
frequently did, ramp down faster than its submitted
ramp rates during the Relevant Period, indicating that
its submitted ramp rates were lower than the Facility’s
operational capabilities.
97
b) The submission of ramp down rates that were lower
than the Facility’s operational capabilities increased the
magnitude of constrained-off CMSC payments to
Bowater.
c) The ramping down of the Facility faster than the
submitted ramp rates increased the magnitude of
constrained-off CMSC payments to Bowater.
11
Failure to
Ramp
The occasions during the Relevant Period where the Thunder
Bay Facility failed to ramp after bidding to do so were
infrequent. While the resulting CMSC payments were selfinduced and should have been clawed back, the available
evidence does not indicate that the failures to ramp were
intentional attempts by Bowater to exploit a market defect.
99
12
Constrainedoff Dispatch
Deviations in
Non-Ramping
Hours
Instances of dispatch deviation in non-ramping hours by the
Thunder Bay Facility during the Relevant Period were
evidence does not indicate that these deviations were
intentional attempts by Bowater to exploit a market defect.
101
13
Profit or
Benefit to
Bowater
Bowater profited $11.0 million from the CMSC payments
received as a result of the behaviours set out in Findings #4
and #8 – 10, which exploited the market defects set out in
Finding #1.
102
14
Expense or
Disadvantage
to the Market
All customers in the wholesale energy market were
disadvantaged by paying additional Uplift charges of
$0.12/MWh as a result of Bowater’s behaviours.
103
15
Finding of
Gaming
Bowater exploited market defects. In so doing, Bowater
received $11.0 million in CMSC payments during the Relevant
Period, and there was a corresponding disadvantage or expense
to the market. Bowater’s conduct constitutes gaming.
104
PUBLIC VERSION
Products Inc.
180
10.2 Findings – Abitibi
The Panel’s findings with respect to its investigation of Abitibi are as follows:
Table 10-2: Summary of Findings Related to Abitibi and the Fort Frances Facility
Page
No.
Subject
Finding
1
Market
Defects
Related to
Constrainedoff CMSC
16
Market
Defects
Related to
Constrainedon CMSC
17
CMSC
Ramping
Strategy
existed during the Relevant Period allowed a dispatchable
load to receive constrained-off CMSC payments that
exceeded the amount required to compensate for reductions
in operating profits arising from responses to dispatch
existed during the Relevant Period allowed a dispatchable
load to receive constrained-on CMSC payments that
exceeded the amount required to compensate for reductions
in operating profits arising from responses to dispatch
Abitibi developed strategies to self-induce CMSC payments
at the Fort Frances Facility, and these were known to senior
management.
18
Knowledge
of CMSC
Compensatio
n Principles
Abitibi was aware that the CMSC regime assumed that
dispatchable loads would bid based on their Marginal
Benefit of Consumption and that CMSC payments were
designed to compensate a dispatchable load for operating
profit reductions when it was directed by the IESO to
follow a dispatch different from its market schedule.
PUBLIC VERSION
118
118
123
124
Products Inc.
181
No.
Subject
19
Operating
Profit
Impact of
Being
Constrained
Off
Page
Finding
a) During periods when Abitibi was not operating the
Fort Frances Facility at capacity, there would be
virtually no reduction in operating profits as a result
of being constrained off during a ramping hour
because production could be made up in a
subsequent hour.
129
b) Even in situations where the Fort Frances Facility
was capacity constrained, Abitibi’s bid prices during
the Relevant Period substantially exceeded its
Marginal Benefit of Consumption and the reduction
in operating profits that would result from the net
load at the Fort Frances Facility being constrained
off during ramping hours.
c) Based on data provided by Abitibi and the Facility’s
electricity consumption pattern, the difference
between Abitibi’s bid price of $●/MWh (or
$●/MWh) and its Marginal Benefit of Consumption
when ramping (up or down) was at least
$1,418/MWh (or $1,369/MWh).
d) Abitibi’s high bid prices were used to obtain CMSC
payments that more than compensated Abitibi for
20
Risk of
Being
Constrained
Off
The risk of being constrained off during self-induced
ramping hours did not justify Abitibi’s use of a bid price of
$●/MWh or $●/MWh, or any other level above the
Marginal Benefit of Consumption of the Fort Frances
Facility.
132
21
Risk of
Being
Activated for
Operating
Reserve
The risk of being activated to provide operating reserve
during self-induced ramping hours did not justify Abitibi’s
use of an energy market bid price of $●/MWh or $●/MWh,
or any other level above the Marginal Benefit of
Consumption of the Fort Frances Facility.
133
22
High Bid
Prices by
Other Loads
The historical use of high bid prices by Abitibi or any other
dispatchable loads does not provide a justification for
Abitibi’s high bid prices during self-induced ramping
hours.
134
PUBLIC VERSION
Products Inc.
182
No.
Subject
23
Ramping
Down Faster
than
Submitted
Rates
Page
Finding
a) Abitibi’s Fort Frances Facility was able to, and
frequently did, ramp down the net load faster than
its submitted ramp rates, indicating that its ramp
rates were lower than its operational capabilities.
137
b) The submission of ramp rates that were lower than
the Fort Frances Facility’s operational capabilities
increased the magnitude of constrained-off CMSC
payments to Abitibi.
c) The ramping down of the Fort Frances Facility
faster than the submitted ramp rates increased the
magnitude of constrained-off CMSC payments to
Abitibi.
24
Frequent
Ramping
Abitibi increased its CMSC payments through frequent
ramping of the Fort Frances Facility during the Relevant
Period by at least $5.8 million.
143
25
Bid Prices
When Using
Generator to
Alter Net
Consumptio
n
When Abitibi used the generator to implement self-induced
changes to the net load at the Fort Frances Facility, the bid
prices it submitted did not reflect the marginal cost of the
generating facility, resulting in CMSC payments that
substantially exceeded the amount needed to compensate
Abitibi for any operating profit reductions, but was not a
deliberate attempt to exploit a market defect.
146
26
Ramp Rates
When Using
Generator to
Alter Net
Consumptio
n
When Abitibi used the generator to implement self-induced
changes to the net load at the Fort Frances Facility, the
ramp rates it submitted did not reflect the actual ramping
capabilities of the generator, resulting in CMSC payments
that substantially exceeded the amount needed to
compensate Abitibi for any operating profit reductions, but
was not a deliberate attempt to exploit a market defect.
147
27
Failure to
Ramp
The occasions during the Relevant Period where the Fort
Frances Facility failed to ramp after bidding to do so were
evidence does not indicate that these failures to ramp were
attempts by Abitibi to exploit a market defect.
149
PUBLIC VERSION
Products Inc.
183
Page
No.
Subject
Finding
28
ConstrainedOff Dispatch
Deviations in
Non-Rampin
g Hours
Instances of dispatch deviation in non-ramping hours by the
Fort Frances Facility were infrequent. While the CMSC
payments were self-induced and should have been clawed
back, the available evidence does not indicate that these
deviations were attempts by Abitibi to exploit a market
defect.
151
29
ConstrainedOn CMSC
Payment
Strategy
Abitibi’s adoption of a negative-price bidding strategy
between April and August 2010, which was known to
senior management, was a deliberate attempt to exploit a
market defect in the CMSC regime that had been publicly
identified as such by the Panel and was in the process of
being rectified by the IESO.
161
30
Negative Bid
Prices
a) Abitibi’s -$●/MWh bid price was well below its
Marginal Benefit of Consumption during the hours
in which such bids were submitted for the net load
at the Fort Frances Facility.
163
b) Abitibi’s low bid price was used to obtain CMSC
payments that more than compensated Abitibi for
31
ConstrainedOn Dispatch
Deviations
Abitibi deviated from the Fort Frances Facility’s dispatch
instructions on numerous occasions that resulted in the
Facility appearing to be constrained on and receiving
CMSC payments when its bids indicated it did not want to
consume.
164
32
Profit or
Benefit to
Abitibi from
Constrainedoff CMSC
Abitibi profited $7.5 million from the constrained-off
CMSC payments received as a result of the behaviours set
out in Findings #19, 23 and 24, which exploited the market
defects set out in Finding #1.
166
33
Profit or
Benefit to
Abitibi from
Constrainedon CMSC
Abitibi profited $1.9 million from the constrained-on
CMSC payments received as a result of the behaviours set
out in Findings #30 and 31, which exploited the market
defects set out in Findings #1 and #16.
168
PUBLIC VERSION
Products Inc.
184
Page
No.
Subject
Finding
34
Expense or
Disadvantag
e to the
Market
All customers in the wholesale energy market were
disadvantaged by paying additional Uplift charges of
$0.09/MWh as a result of Abitibi’s behaviours.
169
35
Finding of
Gaming
Abitibi exploited market defects. In so doing, Abitibi
received $9.4 million in CMSC payments during the
Relevant Period, and there was a corresponding
disadvantage or expense to the market. Abitibi’s conduct
constitutes gaming.
170
10.3 Recommendation
The Panel makes the following recommendation to the IESO:
a) The IESO should review the CMSC payments being made to dispatchable loads since
the November/December 2010 amendments to the Market Rules in order to determine
whether there are significant amounts that continue to be unwarranted (i.e., paid as a
result of market participant actions rather than to compensate for operating profit
reductions arising from responding to dispatch instructions caused by Grid
Conditions).
b) If necessary, the IESO should make further amendments to the Market Rules to
eliminate unwarranted CMSC payments to dispatchable loads.
PUBLIC VERSION
Products Inc.
185
Appendix A Glossary
Term/
Abbreviation
Definition
Page
Act
Electricity Act, 1998
23
ABI
AbitibiBowater Inc. (ultimate parent company of Abitibi and
Bowater)
18
Abitibi
Abitibi-Consolidated Company of Canada, owner and operator
of the Fort Frances Facility
17
Bowater
Bowater Canadian Forest Products Inc., owner and operator of
the Thunder Bay Facility
17
Business Rules
The IESO initially relied on manual processes to identify
CMSC payments that should be recovered. In 2007, the IESO
introduced an automated approach to CMSC recovery, and
documented the procedures used to calculate the amount of
participant-induced CMSC that may be clawed back under the
Market Rules in four Business Rules (see overview in
Appendix H)
40
CCAA
Companies’ Creditors Arrangements Act
18
CMSC
Congestion Management Settlement Credit
9
Compliance
Deadband
The allowable range of variation in actual consumption
relative to the IESO’s dispatch instructions. (The Compliance
Deadband for a resource with the characteristics of the
Thunder Bay Facility or the Fort Frances Facility is 15 MW
above or below its dispatch instruction.)
83
Constrained
Schedule
Energy injections and withdrawals for dispatchable facilities
that can actually happen within the physical constraints of the
system
32
Dispatch
Envelope
The range between the maximum and minimum dispatch
instruction based on the load’s current consumption level and
submitted ramp rates
98
Dispatch
Schedule
See Constrained Schedule
31
PUBLIC VERSION
Products Inc.
186
Term/
Abbreviation
Definition
Page
DR2 Agreement
A 2009 agreement between Bowater and the Ontario Power
Authority which preceded the reopening of the Thunder Bay
Facility
43
Fort Frances
Facility
Abitibi’s facility in Fort Frances, Ontario includes three paper
machines (two of which were active), one kraft mill, and a
biomass and natural gas boiler/generator
19
Grid Conditions
All physical limitations on the grid, including transmission
constraints and transmission line losses
31
h
one hour
46
h+1
the following hour
46
HOEP
Hourly Ontario Energy Price
29
IESO
Independent Electricity System Operator
17
Lamination
One component of a dispatchable load’s bid for electricity
consumption, consisting of a P/Q Pair. A bid may have up to
20 laminations.
30
MAU
The IESO’s Market Assessment Unit
23
Marginal Benefit
of Consumption
The Marginal Benefit of Consumption is the incremental net
revenue expected to result from increasing production by
consuming an additional MW of electricity. “Net revenue” is
the revenue expected to result from selling the additional
output less variable costs of production other than electricity.
A firm normally would not be prepared to pay more than the
Marginal Benefit of Consumption. If it did so, the cost of the
extra MW would exceed the incremental net revenue from
increasing output (i.e., its operating profits would be reduced).
The Marginal Benefit of Consumption may also be used to
measure the lost net revenues (again, before considering
electricity costs) when a firm consumes one less MW of
electricity. A firm normally would not reduce its consumption
if the Marginal Benefit of Consumption exceeded the price of
electricity
37
PUBLIC VERSION
Products Inc.
187
Term/
Abbreviation
Definition
Page
Market defect
A defect in the market design, poorly specified rules or
procedures or a gap in the Market Rules or procedures
27
Market Schedule
See Unconstrained Schedule
24
MCP
Market Clearing Price
29
MMCP
Maximum Market Clearing Price ($2,000/MWh)
30
MSP
17
MSP By-Law
OEB By-Law #3 – Market Surveillance Panel
23
Nodal Price
The system marginal cost of supply at a point on the grid
33
OR
Operating Reserve
78
Operating Profit
For each MW of consumption, the difference between a
dispatchable load’s Marginal Benefit of Consumption for the
MW and the price paid for consuming the MW.
62
Panel
17
P/Q Pair
One lamination of a bid by a dispatchable load, consisting of a
bid price (“P”) and the corresponding quantity (“Q”) that the
load is prepared to consume at that price.
45
Protocol
Protocol between the Ontario Energy Board and the
Independent Electricity System Operator Related to Market
Surveillance Panel
23
Ramp Rate
How quickly a load (or generator) can change (upwards or
downwards) the amount of energy it is consuming (or
producing), expressed in MW/minute
33
Relevant Period
January 2010 to August 2010
17
Responses to
RFI
Responses from Bowater and Abitibi to the Panel’s requests
for information.
26
PUBLIC VERSION
Products Inc.
188
Term/
Abbreviation
Definition
Page
Thunder Bay
Facility
Bowater’s facility in Thunder Bay, Ontario includes a thermomechanical pulpmill with two mainline refiners, a rejects
refiner, auxiliaries, and a recycle mill
18
TMP
Thermo-mechanical pulpmill
18
Unconstrained
Schedule
The amount of energy that dispatchable facilities would be
prepared to inject or withdraw if there were no constraints on
the system
33
Uplift
CMSC payments are charged pro rata to all wholesale loads
(including exports) through hourly uplift charges
11
PUBLIC VERSION
Inc.
189
Appendix B Abitibi Bowater Inc. Corporate Chart
Abitibi Bowater Inc.
(now Resolute Forest Products Inc.)
(“ABI”)
100%
82%
Bowater Incorporated
100%
Bowater Canadian Holdings Incorporated
100%
Abitibi Bowater Canada Inc.
100%
18%
Bowater Canadian Forest Products Inc.
(“Bowater”)
Abitibi-Consolidated Inc.
100%
Abitibi-Consolidated Company of Canada
(subsequently Abitibi Canada Inc, and
now Resolute FP Canada Inc.)
(“Abitibi”)
PUBLIC VERSION
Inc.
190
Appendix C Selected Bowater and Abitibi Personnel who Prepared or Received
Communications and Documents Referred to in this Report
Name
Position
Company
●
●
Bowater
Thunder Bay Operations
●
●
●
●
●
●
●
●
●
●
AbitibiBowater Inc.
Montreal Operations
●
●
Abitibi
Fort Frances Operations
●
●
●
●
●
●
●
●
●
●
●
●
Abitibi
Abitibi
AbitibiBowater Inc.
Montreal Operations
Bowater
Bowater
Abitibi
AbitibiBowater Inc.
Montreal Operations
Bowater
Bowater
AbitibiBowater Inc.
PUBLIC VERSION
Products Inc.
191
Name
Position
Company
●
●
AbitibiBowater Inc.Montreal
Operations
●
●
AbitibiBowater Inc.
Montreal Operations and Sales
●
●
●
●
●
●
●
●
Abitibi
AbitibiBowater Inc.
Montreal Operations
Bowater
Bowater
PUBLIC VERSION
Inc.
192
Appendix D The Unconstrained Mode and Schedule
D.1
Determining the Market Clearing Price and Market Schedules
The Market Clearing Price (MCP) is the price at which the supply of electricity is equal to the
demand for electricity. This is the point where a downward sloping demand curve intersects with
an upwards sloping supply curve. Conceptually, the MCP is established by stacking all offers of
supply from the lowest to the highest offer price until the total quantity offered equals the
amount of electric power demanded.244
The MCP is calculated for each five-minute interval and the MCPs are averaged to calculate the
HOEP. Subject to various adjustments, all wholesale market suppliers are remunerated on the
basis of the MCP or HOEP, even if they offered the energy at a lower price, and all wholesale
market customers pay the MCP or HOEP, even if their bid indicated a willingness to consume at
a higher price.245
There are three particularly important characteristics which distinguish the unconstrained mode
from the constrained mode: absence of physical constraints; a ramp-rate-multiplier; and
single-interval optimization. Each of these characteristics cause the quantities determined in the
unconstrained schedule to differ from those in the constrained schedule.
D.2
Absence of Physical Constraints
The unconstrained mode basically ignores transmission constraints inside Ontario. This allows
the calculation of a five-minute MCP (or HOEP) that is the same for all energy market
participants in Ontario because it does not consider line losses, transmission congestion, and
other system constraints that would otherwise cause supply/demand conditions and prices to
differ from location to location on the grid.
244
Electric power demanded is the sum total of fixed electric power demanded from non-dispatchable loads and
exports plus the amount of variable electric power demanded from dispatchable loads that are priced higher than any
offers of supply that remain after the demand from non-dispatchable load and exports has been satisfied.
245
Dispatchable generators receive, and dispatchable loads pay, the five-minute MCP. Non-dispatchable generators
and loads are settled using HOEP (as are importers and exporters, subject to localized congestion price adjustments).
PUBLIC VERSION
Products Inc.
193
D.3
Ramp-Rate-Multiplier
The unconstrained mode unrealistically assumes dispatchable facilities can ramp up or down
three times faster than the ramp rates they submit to the IESO. The original Ontario market
design did not include such an assumption. However, in market testing before the market
opening in May 2002, the IESO discovered that market prices could be volatile during periods
when the whole system was ramping up or down. The Market Rules were amended just before
market opening to require that the unconstrained mode use ramp rates that were 12 times faster
than the ramp rates submitted by market participants.246 This assumption was changed to a
three-times ramp-rate-multiplier in September 2007 after a proceeding before the OEB.247 The
ramp-rate-multiplier means that market schedules for dispatchable facilities will, in periods
where the facility is ramping up or down, include quantities that the facility could never generate
or consume.
D.4
Single-Interval Optimization
In the unconstrained mode, the dispatch algorithm looks backwards to the interval that just ended
as a starting point to determine the MCP and market schedule quantities for the current interval.
The results for each interval are calculated in isolation from all other intervals (i.e., the dispatch
algorithm optimizes the market schedule for a single interval without consideration of any future
intervals). As a result, the market schedule can change significantly from one interval to the next
as the dispatch algorithm reacts to changes in generator and importer (supply) offers,
dispatchable load and exporter (demand) bids, and estimated demand from non-dispatchable
loads. Typically, such changes are largest “across-the-hour” as the unconstrained mode considers
a set of hourly offers and bids in interval 1 of the new hour that may be significantly different
from those applicable during the prior hour.
246
IESO, Market Rule Amendment MR-00189, April 18, 2002, online (as an attachment to Market Pricing Working
Group Memorandum): http://www.ieso.ca/imoweb/pubs/consult/mep2/MP_WG-200708023Memo_Action_Item_43-1.pdf, p. 3.
247
IESO, Market Rule Amendment MR-00331, September 12, 2007, online:
http://www.ieso.ca/Documents/Amend/mr2007/MR-00331-R00-BA.pdf; and OEB, Decision Order, EB-2007-0040,
April 12, 2007, online: http://www.ontarioenergyboard.ca/documents/cases/EB-20070040/dec_order_revised_ampco_20070412.pdf, p. 26.
PUBLIC VERSION
Products Inc.
194
D.5
The Market Schedule
The unconstrained mode determines the market schedule for each dispatchable market
participant based on the absence of physical constraints, the ramp-rate-multiplier and
single-interval optimization. The quantities shown in market schedules are not used to calculate
the basic energy payments or charges to market participants: they get paid or charged based on
actual injection or withdrawal quantities. However, the MCPs are used as the starting point for
settlement calculations and the quantities in the market schedules are used for determining
CMSC payments (see Appendix F).
PUBLIC VERSION
Inc.
195
Appendix E The Constrained Mode and Schedule
E.1
Determining Dispatch Instructions and Nodal Prices
While the unconstrained mode focuses solely on economics, the constrained mode considers
both economics and system limitations. There are three particularly important characteristics
that are taken into account during the constrained mode: physical constraints; actual ramp rates;
and multi-interval optimization. Each of these characteristics cause the quantities determined in
the constrained schedule to differ from the unconstrained schedule.
The constrained mode of the dispatch algorithm also calculates Nodal Prices at each physical
location on the transmission system where energy is injected by generators or withdrawn by
loads. If suppliers’ offers reflect their marginal cost of supplying electricity, and customers’ bids
reflect the marginal value they place on consumption, then the Nodal Prices will represent the
locational value of the energy (including the cost of any line losses, and the impact of
congestion) at each node.
E.2
Physical Constraints
Initially, the constrained mode stacks offers and bids economically. It then determines whether
or not it can dispatch the facilities in economic order and still respect system limitations such as
line losses and transmission limitations applicable to each of the injection and withdrawal
locations on the grid.
Physical dispatch instructions for a facility are based on the relationship of the facility’s offers or
bids to the Nodal Price determined for its node. For example, the constrained schedule for a
dispatchable load will not schedule any energy withdrawals where the load’s bid price (which
applies throughout the hour) is less than the relevant five-minute interval Nodal Price (even
though the load’s bid price may be well above the five-minute uniform MCP and the
unconstrained mode has included the load as consuming in the market schedule, in which case a
CMSC payment would be calculated).
PUBLIC VERSION
Products Inc.
196
E.3
Actual Ramp Rates
The constrained mode uses the actual ramp rates submitted by market participants, not the
three-times faster ramp-rate-multiplier used in the unconstrained schedule (see Appendix D).
E.4
Multi-Interval Optimization
In the constrained mode, the dispatch algorithm determines dispatch instructions for the next
five-minute interval having regard to expected future system conditions. This multi-interval
optimization (MIO) process considers both economics and system limitations over a number of
intervals, rather than the singleinterval retrospective approach used in the unconstrained schedule
(see Appendix D). With the benefit of foresight, the dispatch algorithm produces a more efficient
dispatch pattern because it recognizes ramp rate limitations, expected changes in
non-dispatchable demand and across-the-hour offer or bid changes in future intervals (based on
hourly pre-dispatch offer and bid submissions). For example, it may ramp slower-moving but
lower-cost dispatchable facilities in advance of the intervals when they are most needed (instead
of more expensive resources with faster ramping capability).248
E.5
The Constrained Schedule
The constrained mode produces a dispatch schedule which identifies the expected supply or
consumption by each dispatchable facility for each five-minute interval. However, the quantities
and the Nodal Prices in the dispatch schedule, are not used as the basis for settlement
calculations (except to the extent that they affect CMSC calculations (see Appendix F)).
248
For more detail, see IESO, Quick Takes 13: Multi-Interval Optimization: An IESO Marketplace Training
Publication, online: http://www.ieso.ca/imoweb/pubs/training/QT13_MIO.pdf.
PUBLIC VERSION
Inc.
197
Appendix F Calculation of Constrained-Off and Constrained-On CMSC Payments
The CMSC payment for a dispatchable load in any five-minute interval is effectively calculated
as the difference between the bid price and MCP, multiplied by the difference between
unconstrained schedule and the constrained schedule (or in certain circumstances the load’s
actual consumption). More precisely:249
Constrained-off CMSC = [Bid price – MCP] x [MQSW – max (DQSW, AQEW)]
Constrained-on CMSC = [MCP – Bid price] x [min (DQSW, AQEW) – MQSW]
Where:
MQSW is the quantity in the load’s market (unconstrained) schedule,
DQSW is the quantity in the load’s dispatch (constrained) schedule, and
AQEW is the quantity actually consumed by the load during the interval.
There is an important distinction between DQSW and AQEW. DQSW is a MW quantity that the
IESO instructs a dispatchable facility to meet by the end of the next five-minute interval. AQEW
is the actual energy withdrawn by the load over the entire five-minute interval as measured by
the facility’s revenue meter. Where the load varies its consumption over the course of an
interval, the reported AQEW value will be an average value. Figure F-1 shows how the DQSW
and AQEW differ when a load receives a series of decreasing dispatch instructions during a ramp
down and it fully complies with dispatch instructions.
249
The last parts of these simplified equations are expressed as either a maximum or minimum of two quantities.
This is necessary to ensure CMSC will not be paid when a load does not fully adjust its consumption to the level
required by the IESO’s dispatch instructions. See Market Rules, ch. 9, s. 3.5 for the precise CMSC equations.
PUBLIC VERSION
Products Inc.
198
Figure F-1: Illustration of How CMSC is Calculated Using the Maximum of AQEW and
DQSW During a Ramp Down
PUBLIC VERSION
Inc.
199
Appendix G CMSC Payments Arising from Self-Induced Ramping
Ramping decisions made by a market participant may result in self-induced CMSC payments.
More specifically, changes in the bid (or offer) quantities and/or prices of a market participant
can self-induce ramping of a facility and give rise to CMSC payments that are not caused by
IESO actions related to transmission or security limitations and that may overcompensate a load
(or generator).
G.1
Ramp-Rate-Multiplier on Ramp Ups
The three-times ramp-rate-multiplier that is built into a load’s unconstrained schedule can
contribute to CMSC payment through differences between the market and dispatch schedules
whenever a load takes more than one interval to ramp up or down to a desired level of
consumption. Depending on the load’s ramp rates, the quantities in the unconstrained and
constrained schedules may deviate during several intervals. All else being equal, the slower a
load’s ramp rate, the greater the deviation between the unconstrained and constrained schedules,
and the higher the CMSC payments will be. This type of CMSC arises on ramps induced by a
change in a facility’s bids, as well as dispatches arising from changes in market supply/demand
conditions.
Figure G-1 illustrates how a load’s ramp rates affect schedule quantity differences (and hence the
amount of CMSC) when the load is ramping. Assume the facility is ramped up from 10 MW to
100 MW. The solid lines represent the ramp in the constrained sequence: the blue line indicates a
fast ramp rate and the red line, a slower ramp rate. The solid lines represent the fast or slow
ramps in the constrained schedule and the dashed diagonal lines represent the unconstrained
schedule (based on the three-times ramp-rate-multiplier). The quantity difference (and hence
CMSC payment) is relatively small where the load ramps up quickly (blue shaded area).
However, where the load ramps up slowly, a larger quantity difference (and CMSC payment) is
generated (red shaded area).
PUBLIC VERSION
Products Inc.
200
Figure G-1: Illustration of How CMSC May be Increased Using a Lower Ramp Rate During
Self-Induced Ramp Up
G.2
Ramp-Rate-Multiplier on Ramp Downs
Figure G-2 illustrates how a load’s ramp rates affect schedule quantity differences (and hence the
amount of CMSC) when the load self-induces a ramp down. Assume the facility is ramped down
from 100 MW to 10 MW. The quantity difference (and hence CMSC payment) is relatively
small where the load ramps down quickly (blue shaded area). However, where the load ramps
down slowly, a larger quantity difference (and CMSC payment) is generated (red shaded area).
PUBLIC VERSION
Products Inc.
201
Figure G-2: Illustration of How CMSC May be Increased Using a Lower Ramp Rate During
Self-Induced Ramp Down
G.3
Single and Multi-Interval Optimization
Single versus multi-interval optimization causes similar variations between constrained and
unconstrained schedules, again leading to CMSC payments. Under MIO, the constrained
schedule is set on a “looking forward” basis, whereas the unconstrained schedule looks
backward. When MIO looks forward for an upcoming hour ‘h+1’and sees that the quantity bid
has been reduced, the IESO dispatch tool begins ramping down the constrained schedule in
advance of the next hour to ensure the facility is consuming no more than its quantity bid by
interval 1 of hour ‘h+1’. In contrast, the unconstrained schedule does not look ahead to hour
‘h+1’, it only sees the full quantity bid in the current hour (i.e. it does not consider that the bid
quantity has dropped in hour ‘h+1’ until the beginning of hour ‘h+1’ when the schedule starts
ramping down). As the load ramps down in advance of hour ‘h+1’ the market and dispatch
schedules will diverge. As such, the load appears to be constrained off and receives CMSC, even
though the facility was being dispatched in accordance with its quantity bids and was consuming
no more and no less than it desired.
PUBLIC VERSION
Inc.
202
Appendix H IESO “Business Rules” for Clawback of Constrained-Off CMSC
Payments
In 2007, the IESO introduced an automated approach to CMSC recovery. It also documented the
procedures used that would be to calculate the amount of participant-induced CMSC that may be
clawed back under the Market Rules. The four criteria — referred to as “Business Rules” —
which are applied by the IESO to recover constrained-off CMSC from dispatchable load, are
summarized below:250
H.1
Business Rule 1 – Materiality
Constrained-off CMSC is allowed for an interval if the total amount of CMSC paid during that
trading day to that dispatchable load is less than $4,000.
H.2
Business Rule 2 – Non-Dispatchable Portion of Load
Constrained-off CMSC is not allowed for an interval if it is paid for portions of the schedule
where the load has bid at +MMCP (i.e. $2,000/MWh), indicating that it is non-dispatchable in
that quantity range.
H.3
Business Rule 3 – Dispatch Deviation
Constrained-off CMSC is not allowed for an interval if the current 5-minute constrained
schedule exceeds the revenue meter value (i.e., actual consumption) in the previous interval plus
2.5 minutes of ramping. However, this rule does not apply in various circumstances, including:
when the load is constrained-off economically; when the load is ramping; and when the load is
manually dispatched down for reliability.
250
IESO, Market Manual 5, Part 5.5: Physical Markets Settlement Statements, s. 1.6.9. The Market Rules and IESO
procedures do not address recovery of constrained-on CMSC.
PUBLIC VERSION
Products Inc.
203
H.4
Business Rule 4 – Facility Off-Line or Unable to Follow Dispatch
Constrained-off CMSC is not allowed for an interval if the constrained schedule is 0 MW and
consumption is less than 1 MW, or if consumption is 0 MW. However, this rule does not apply
in various circumstances, including: when the load is constrained-off economically; and when
the load is manually dispatched down for reliability.
H.5
Intervals
There are over 100,000 five-minute intervals in a year, and there will be differences in many of
those intervals between the constrained and unconstrained schedules for dispatchable loads and
generators. It is a tall order to search out and clawback all inappropriate CMSC payments when,
as described in Section 6 and Appendix D through Appendix G, there are various reasons why
the quantity schedules can differ.
PUBLIC VERSION
Inc.
204
Appendix I
Five Largest CMSC Payment Days for Bowater’s Thunder Bay Facility
Table I-1 below shows the five days on which Bowater received the highest CMSC payments for
the Thunder Bay Facility in 2010. On each of these days the CMSC payments received far
exceeded the charges incurred for the energy consumed.
Table I-1: Highest CMSC Daily Payments and Net Energy Cost for the Thunder Bay Facility
January to August 2010
($000)
Date
April 11, 2010 (Sunday)
March 05, 2010 (Friday)
February 21, 2010 (Sunday)
May 16, 2010 (Sunday)
April 26, 2010 (Monday)
Total Top Five
Energy
Charges*
CMSC
Payments**
35
89
133
127
88
$471
Net Energy
Costs
(Revenue)
215
179
153
128
123
$798
(180)
(90)
(20)
(1)
(35)
$(326)
** All amounts are net constrained-off CMSC payments after clawback of certain types of CMSC (charge
type 105 less charge type 1050).
Figures I-1 to I-5 below show the hourly constrained and unconstrained schedules, as well as
actual consumption and CMSC payments, on each of the five days listed above.
PUBLIC VERSION
Products Inc.
205
Figure I-1: Schedules and CMSC Payments for the Thunder Bay Facility
April 11, 2010
($ and MW)
March 5, 2010
($ and MW)
PUBLIC VERSION
Products Inc.
206
February 21, 2010
($ and MW)
May 16, 2010
($ and MW)
PUBLIC VERSION
Products Inc.
207
April 26, 2010
($ and MW)
PUBLIC VERSION
Inc.
208
Appendix J
AbitibiBowater Canada Inc. Presentation Slides Discussing CMSC and
Operating Reserve
An internal PowerPoint presentation prepared by [Senior Bowater Personnel #2] for Abitibi
Bowater’s Vice Presidents, entitled “Thunder Bay 2010 Power Cost – October 1st, 2009”,
provided a summary of various financial programs, including CMSC and Operating Reserve.251
Four slides from the presentation which discuss CMSC and Operating Reserve are reproduced
below.
Slide Redacted – Contains Confidential Information
251
PUBLIC VERSION
Products Inc.
209
PUBLIC VERSION
Products Inc.
210
PUBLIC VERSION
Inc.
211
Appendix K Sample Bowater Calculations for the Thunder Bay Facility
Relating to Ramping Down Faster than Submitted Ramp Rates
Below is a spreadsheet prepared by personnel at the Thunder Bay Facility which shows how
Bowater planned and forecasted their “Optimized CMSC Payment (actual load < constrained
schedule)”, which entails ramping faster than submitted ramp rates.252
Spreadsheet Redacted – Contains Confidential Information
252
Attachment titled “TMP Shutdown & Stat Up Sequence re Dispatchable.xls” in email from [Senior Bowater
Personnel #5] to [Senior AbitibiBowater Inc Personnel #2] and [Senior Abitibi Personnel #2], January 22, 2010.
PUBLIC VERSION
Inc.
212
Appendix L Sample Bowater Analysis of CMSC During Ramp Down
Below is a chart prepared by personnel at the Thunder Bay Facility which shows Bowater
understood that the CMSC calculation was based on the MW difference between the
unconstrained schedule and the maximum of the constrained schedule or the actual energy
withdrawn. 253 The chart and sample CMSC calculation shows the Thunder Bay Facility’s
consumption as greater than the constrained dispatch instruction during the first two ramp down
intervals.
Chart Redacted – Contains Confidential Information
253
Included in email from [Senior Bowater Personnel #5] to [Senior Bowater Personnel #3], September 25, 2009,
PUBLIC VERSION
Inc.
213
Appendix M Five Largest CMSC Payment Days for Abitibi’s Fort Frances Facility
Table M-1 below shows the five days with the highest CMSC payments for the Fort Frances
Facility between January and August 2010. On each of these days the CMSC payments received
far exceeded the charges incurred for the energy consumed by the net load.
Table M-1: Highest CMSC Daily Payments and Net Energy Cost
for the Fort Frances Facility
January to August 2010
($000)
Date
Energy
Charges*
CMSC Payments*
Constrained-Off
June 1 (Tuesday)
July 21 (Wednesday)
July 22 (Thursday)
April 16 (Friday)
June 2 (Wednesday)
Total Top Five
Constrained-On
Scenario #1
Net Energy
Costs
Constrained-On
Scenario #2
Total
48
24
27
27
40
1
63
5
52
50
164
4
42
210
98
174
220
222
-3
84
339
287
269
259
233
(291)
(263)
(242)
(231)
(193)
166
171
518
697
1,386
(1,220)
** All amounts are net CMSC payments after clawback of certain types of CMSC payments (charge type 105 less
charge type 1050).
Figures M-1 to M-5 below show the hourly constrained and unconstrained schedules, as well as
actual consumption and CMSC payments, on each of the five days listed above.
PUBLIC VERSION
Products Inc.
214
Figure M-1: Schedules and CMSC Payments for the Net Load at the Fort Frances Facility
June 1, 2010
(MW and $)
July 21, 2010
(MW and $)
PUBLIC VERSION
Products Inc.
215
July 22, 2010
(MW and $)
April 16, 2010
(MW and $)
PUBLIC VERSION
Products Inc.
216
June 2, 2010
(MW and $)
PUBLIC VERSION
Products Inc.
217
Appendix N Resolute’s July 2, 2014 Response
PUBLIC VERSION
Products Inc.
247
Appendix O The Panel’s Comments on Resolute’s July 2, 2014 Response and Update on
Subsequent Correspondence and the General Conduct Rule
In accordance with section 7.2.2 of the MSP By-Law, the Panel provided a draft of this Report to
the market participants on April 16, 2014, to provide them with an opportunity to discuss the
findings with the Panel, to respond to the findings and to comment on matters of factual accuracy
and confidentiality. The Panel offered to meet with the market participants, and identified the
date by which any written response should be provided. Resolute delivered a written response to
the Panel’s draft report on July 2, 2014, which is reproduced in Appendix N. Appendix N as it
appears in the public version of this Report was redacted by Resolute.
While Resolute’s response directly addresses some of the Panel’s findings, it more generally
attacks the integrity of the Panel’s process, including allegations that the Panel has acted in a
manner that is biased and unfair. The Panel deals first with these latter issues, and then turns to
Resolute’s response on the substance of the Panel’s findings.
A.
The Panel’s Comments on Resolute’s Claims regarding the Integrity of the Panel’s
Process and Similar Issues
1.
Claims of bias and unfair process
Resolute’s response alleges that the Panel is “biased in its analysis and conclusions.” The Panel
has carefully considered each claim made in Resolute’s response to the effect that the Panel has
not fairly considered or fairly characterized the materials before it, and believes those claims to
be without substance.
Resolute provides six examples of materials that it alleges were “ignored” or “distorted” by the
Panel (these are found at pages 15 through 17 of Resolute’s response). Some of these examples
are specifically addressed elsewhere below, and in the Panel’s view none provide any real basis
for Resolute’s claims. To illustrate, the Panel refers to Resolute’s sixth example, found at page
17 of its response, in which Resolute claims that the Panel’s use of $●/MWh as the estimated
Marginal Benefit of Consumption for the Thunder Bay Facility – as opposed to the $●/MWh
PUBLIC VERSION
Products Inc.
248
advocated by Resolute – had as its purpose to inflate the magnitude of Resolute’s unwarranted
CMSC payments. Resolute also states that the $●/MWh was referred to on only one occasion in
the e-mails provided to the Panel, and that “as soon as the $● is raised in the internal
correspondence, it is clearly identified as an error.”
The Panel’s decision to use $●/MWh was made after careful review of both that figure and the
$●/MWh figure advanced by Resolute, including all relevant materials provided by Resolute on
this point. E-mails and associated e-mail attachments related to the internal correspondence
referred to by Resolute regarding the erroneous identification of the $●/MWh figure indicate that
additional fixed costs and downtime of the paper machine were added to derive the $●/MWh
figure. As explained in section 7.4.2.2 of this Report, the Panel believes these costs to be
inappropriate in calculating the marginal benefit. The Panel’s conclusion is that, for the reasons
discussed in sections 7.4.2.2 to 7.4.2.4 of this Report, $●/MWh overstates the marginal benefit.
The Panel is not required to accept a market participant’s own estimate of the marginal benefit,
and is not precluded from using an alternative estimate in cases such as this where the materials
before the Panel give it reason to do so. Although the Panel in fact believes that even the
$●/MWh amount is too high, as noted in this Report the Panel has used that figure in the absence
of any better information.
Resolute also complains that the Panel’s process is “fundamentally unfair” because there is no
requirement for the Panel to prove its allegations before an independent tribunal. In effect,
Resolute takes the position that any investigation by the Panel will be unfair in the absence of an
associated adjudicative process. The fact that there is no provision for adjudication in the
context of the Panel’s investigations is a decision of the legislature. The Panel’s responsibility is
to ensure that its process accords with its legislated mandate and with the MSP By-law, and the
Panel has done so in this case. It is worth noting in this context that, while invited by the Panel
to do so, Resolute made no attempt to meet with the Panel following receipt of the Panel’s draft
Report.
In reaching its conclusions in this case, the Panel acted carefully and fairly, and believes that its
conclusions will withstand any level of serious scrutiny.
PUBLIC VERSION
Products Inc.
249
2.
Claims that the Panel does not understand the electricity market and is trying to redefine
the rules
Resolute alleges variously that the Panel is “out of touch”, “appears to be… unaware of how the
IESO-administered market actually works” and is “attempting to change the meaning of the rules
on a retroactive and selective basis”. The Panel includes in this general category of claims the
claim that the market defect at issue is not entirely clear.
The Panel is well aware of how the Ontario electricity market works, and is confident that even a
cursory reading of this Report will satisfy an objective reader that that is the case. With respect
to bid prices, as discussed in further detail in section B below the Panel is in no way attempting
to redefine the Market Rules.
Contrary to Resolute’s assertion, the nature of the market defect at issue in this case is entirely
clear from the Panel’s Report. It is captured in the Panel’s Finding #1, and relates exclusively to
the rules and procedures relating to CMSC payments. Moreover, Resolute attributes to the
Panel’s analysis a market defect of its own creation that is not expressed anywhere in this
Report; namely, the “lack of an obligation on market participants to bid at marginal cost.” While
the Panel accepts that an obligation to bid at marginal cost would go a long way towards
eliminating the potential for gaming of CMSC payments given how these payments are
calculated, the Panel has not identified the lack of such an obligation as a market defect, nor does
it equate that with the market defect captured in its Finding #1, nor does it believe that its
approach in essence has the effect of imposing such an obligation.
Resolute’s response also makes it appear that the Panel is unacquainted with dispatch risk, that
is, the risk that a dispatchable load’s facilities could be dispatched down or off by the IESO if the
Nodal Price at its location rises above its bid price, or the risk that the load could be activated to
provide operating reserve. The Panel’s extensive analyses of dispatch risk are described in
sections 7.4.2.6, 7.4.2.7, 8.4.2.5, and 8.4.2.6 of this Report, and its conclusions are clear that
neither the risk of being constrained off nor the risk of being activated for operating reserve
justified the high bid prices used by Abitibi and Bowater.
PUBLIC VERSION
Products Inc.
250
Taken to their logical conclusion, Resolute’s assertions would appear to preclude the Panel from
scrutinizing the bids that underlie CMSC payments simply because the Market Rules do not
require dispatchable loads to bid at their marginal benefit of consumption, even when there is
evidence that the bids have been set at levels designed to maximize CMSC payments and are
much higher than necessary to address any dispatch or other risks. The Panel has a very different
view of this issue. Fundamentally, the Panel does not accept that gaming behaviour is outside of
its purview because the conduct in question is not prohibited by the Market Rules or may be
technically compliant with the Market Rules. That this is the Panel’s view is illustrated by its
Monitoring Document “Generator Prices Used to Signal an Intention to Come Offline”, which
makes it clear that offer prices that are allowable under the Market Rules (because they are no
higher than the maximum market clearing price of $2,000/MWh) can nevertheless be the subject
of a gaming investigation.
3.
Resolute’s suggestion that its behavior was encouraged by IESO staff
In its response, Resolute makes references to discussions or meetings with IESO staff. Through
these references, it appears that Resolute is trying to create the impression that IESO staff
condoned, if not encouraged, the behaviours that resulted in the substantial CMSC payments that
are the subject of this Report.
The Panel makes two observations in this regard. First, the IESO staff members referred to by
Resolute did not have authority to determine what conduct would or would not constitute
gaming, a point that is made in section 8.5.2 of this Report. Second, even if the IESO staff
members had such authority (which again they did not), there is no evidence that any advice that
they might have provided to Abitibi or Bowater was provided with full knowledge of the specific
underlying details of the market participants’ conduct (for example, that the market participants’
bid prices were much higher than required to deal with dispatch risk and would lead to
substantial unwarranted CMSC payments). Merely stating the obvious, such as the fact that high
bid prices lower the risk of being activated to provide operating reserve, is not the same as
encouraging, or condoning, behavior that is undertaken to exploit a defect in the Market Rules.
PUBLIC VERSION
Products Inc.
251
Ultimately, market participants are responsible for their actions in or affecting the wholesale
electricity market, including in respect of conduct that may be subject to review by the Panel for
gaming. The actions of IESO staff referred to in Resolute’s response do not justify or excuse the
market participant behaviours addressed in this Report, any more than did the actions of [Senior
IESO Personnel] as discussed in section 8.5.2 of this Report. In fact, as discussed in more detail
in Section 9 of this Report and in section B below, when the IESO became aware of the situation,
it moved immediately to change the Market Rules to prevent further CMSC payments of the kind
obtained by Abitibi and Bowater through ramping behaviour.
B.
The Panel’s Comments on Resolute’s Response to the Substance of the Panel’s
Findings
In its response, Resolute groups the Panel’s findings into three categories, and the Panel’s
comments are organized accordingly.
1.
Category 1: “Self-Induced CMSC Allegations”
In the first category, Resolute includes Finding #9 (which relates to Bowater’s Thunder Bay
Facility) and Findings #30 and #31 (which relate to Abitibi’s Fort Frances Facility). Resolute
characterizes these as “Self-Induced CMSC Allegations”, and its discussion of them pertains
almost exclusively to CMSC payments that are made in circumstances where a facility fails to
follow dispatch instructions. Resolute makes the point that it has always acknowledged that such
payments should be clawed back, and that by its estimation roughly $5.3 million should have
been returned to the IESO. Of that amount, $1.825 million has been repaid, while the remainder
has not been paid by reason of the IESO’s failure to confirm the amount.
The Panel notes first that, to the extent that Resolute’s descriptor for Findings #9, #30 and #31
(“Self-Induced CMSC Allegations”) reflects a view that CMSC payments can be self-induced
only when there is a failure to follow dispatch instructions, that view is not aligned with the
Panel’s approach. In the Panel’s view, all of the CMSC payments that the Panel finds were
obtained as a result of gaming as set out in this Report were self-induced.
PUBLIC VERSION
Products Inc.
252
In addition, Findings #9 and #30 do not relate to a failure to follow dispatch instructions. Rather,
they relate to ramp down timing and negative bid prices, respectively. Moreover, Resolute’s
response with respect to Finding #9 is not consistent. On the one hand, Resolute includes
Finding #9 in its discussion of the kinds of CMSC payments that it indicates it has been prepared
to repay to the IESO. On the other hand, Resolute also includes that Finding in the third
category in its response, where it defends the timing of the Thunder Bay Facility ramp down and
does not offer to repay any CMSC amounts.
Resolute’s calculation of the CMSC payments that should be repaid in relation to these three
findings ($5.385 million, of which $1.825 has been repaid) differs from the amounts calculated
by the Panel ($7.624 million). If, as Resolute states, the CMSC payments related to Findings #9,
#30, and #31 should be repaid, the Panel encourages Resolute to do so.
2.
Category 2: “Operating Profit Allegations”
The second category of Resolute’s response concerns the Panel’s Findings #1, #4 and #19.
Resolute characterizes this category as “Operating Profit Allegations.” The findings in question
relate to unwarranted CMSC payments made to Resolute resulting from high bid prices during
voluntary ramping activity.
Resolute states, and relies on the fact, that the basis for calculating CMSC payments is a market
participant’s offer or bid prices, not the participant’s marginal cost of production or Marginal
Benefit of Consumption (referred to below as “opportunity cost”). Resolute further states that it
had no obligation under the Market Rules to bid its opportunity costs, and that the bids it
submitted during ramping ($●/MWh) were intended to avoid dispatch risk rather than to exploit
the market rules relating to the calculation of CMSC.
The Panel agrees that there is no section of the Market Rules that states that a market participant
must bid its opportunity costs, but as noted above disagrees that its approach in essence has the
effect of imposing such a requirement. Rather, the Panel’s point is that CMSC payments are
calculated based on the assumption that a market participant’s bids or offers reflect their
opportunity costs, and when this assumption does not hold true the market participant can obtain
PUBLIC VERSION
Products Inc.
253
CMSC payments that go beyond the intended compensatory purpose and that are not always
subject to claw-back by the IESO. This is precisely the market defect identified by the Panel in
Finding #1, and which the Panel concluded was exploited by the market participants in relation
to Findings #4, and #19. Specifically, the Panel concluded that Abitibi and Bowater exploited
this defect while engaged in ramping by consistently submitting an extremely high bid price.
The Panel was not persuaded that these bid prices could be explained as bearing a relationship to
the market participants’ cost of doing business.
Resolute takes the position that its very high bid prices were intended to avoid dispatch risk.
This explanation was also given during the course of the Panel’s investigation. Contrary to the
suggestion in Resolute’s response that this explanation was disregarded by the Panel, the fact
that the Panel considered it is clear from sections 7.4.2.6 and 8.4.2.5 of this Report.
At page 16 of its response, Resolute refers to a slide that it states clearly says that the CMSC
payment projection in the slide is not an attempt to inflate profits, noting that the slide refers to a
CMSC payment as “a side benefit from participating in the OR market and from shutting down
and starting up every day for DR2.” It is not clear to the Panel what difference this distinction
makes, since the “side benefit” would have contributed to profitability in the same way as other
revenue. Moreover, both the Market Design Committee and the Panel have explicitly stated in
the past that Market Rules need to be developed or reformed to deter gaming of “side payments”
of precisely the kind obtained by Resolute in this case.
At page 16 of its response, Resolute asserts that the Panel states that the fact that Bowater did not
“book” all CMSC revenues demonstrates that Bowater’s conduct was intentionally aimed at
increasing CMSC payments. Resolute then claims that the Panel completely ignored materials
that show that Resolute “did not book CMSC revenues from self-induced CMSC because…it
expected that such revenues would be clawed back from the IESO.” In section 7.4.1 of this
Report, the Panel notes that Bowater did not “book” certain CMSC revenues. It does so to
illustrate that Bowater and ABI senior management were aware of the accounting technique.
Moreover, the Panel specifically acknowledges that this accounting treatment was used as a
result of personnel at Bowater and its affiliates being aware that the IESO might seek to recover
PUBLIC VERSION
Products Inc.
254
some of the CMSC payments obtained as a result of their ramping strategies, and that Bowater
only booked what it considered to be “legitimate” CMSC payments.
Finally, the Panel considers it important to complete the record in relation to Resolute’s
description of actions taken and not taken by the IESO in 2010 (page 12 of Resolute’s response).
Resolute states, in the context of Resolute’s argument that the lack of an obligation to bid at
marginal cost is not a defect in the Market Rules, that the IESO expressly considered whether it
should require market participants to bid their marginal costs by means of substituting a market
participant’s bid with a “replacement bid.” Resolute then states that the IESO declined to adopt
this approach on a permanent basis. These statements are made citing the IESO’s Stakeholder
Engagement SE-89, and are accurate. However, to the extent that Resolute invites any inference
to be drawn from these statements, the Panel believes that it is important to note that the solution
that the IESO did adopt – notably, the elimination of constrained-off CMSC payments to
dispatchable loads related to ramping – made the replacement bid approach moot. As described
in section 9.1 of this Report, after the Panel first reported on these issues in its August 2010
Monitoring Report, the IESO took immediate action to temporarily suspend constrained-off
CMSC payments to dispatchable loads altogether in order to “eliminate CMSC payments that are
not consistent with the intent of CMSC payments under the market rules.” It is also clear that
this measure was taken in specific response to CMSC payments that had been made to Bowater
and Abitibi: the IESO noted at the time that “[f]or the period February 1 to July 31 2010, two
market participants received approximately $22 million in CMSC payments associated with two
dispatchable load facilities”; that “[t]hese two dispatchable load facilities represent 22% of
Ontario’s dispatchable load capability yet have received over 95% of the total constrained off
CMSC paid to all dispatchable loads in Ontario”; and that “constrained off CMSC payments
associated with these two facilities (totalling 190 MW) are also equivalent to approximately 75%
of total constrained off CMSC payments made to all dispatchable generators in Ontario during
the same period (approximately 35,000 MW ).”254 The IESO also noted that the main
254
See Urgent Market Rule Amendment Proposal MR-00373-R00, online:
http://www.ieso.ca/Documents/mr/MR_00373-R00.pdf.
PUBLIC VERSION
Products Inc.
255
contributing factor was “frequent ramping up/down with a slow ramp rate and high bid price.”255
Stakeholder Engagement SE-89 was initiated to consider alternative solutions to the temporary
suspension of constrained-off CMSC payments to dispatchable loads, and ultimately culminated
in the Market Rule amendments that permanently eliminated constrained-off CMSC payments to
dispatchable loads associated with self-induced ramping.
3.
Category 3: “Miscellaneous Allegations”
This category of Resolute’s response covers five of the Panel’s findings.
a.
Finding #8 (increase in the maximum bid quantity for the Thunder Bay Facility)
This finding relates to Bowater’s decision to increase the bid quantity for the Thunder Bay
Facility from ● MW to ● MW for the period February 19, 2010 through May 11, 2010.
Resolute claims that the normal range within which the Thunder Bay Facility was intended to
operate was ● MW to ● MW, and that operating in that range depended on an Advanced Quality
Control (“AQC”) system being installed, which had not yet happened when the Thunder Bay
Facility re-entered the market as a dispatchable load. The Panel has seen no documents that
support that the normal operating range for the Thunder Bay Facility was ● MW to ● MW, nor
does the e-mail cited by Resolute in its response suggest that such an operating range was
dependent on installation of an AQC system. In fact, the Panel notes that the consumption level
at the Thunder Bay Facility in off-peak hours averaged less than ● MW for April 2010,
notwithstanding that the AQC system referred to in the response was not in place at that time
(Resolute states that it was implemented in May 2010).
Resolute also claims that the increase in the bid quantity was implemented to avoid noncompliance with the IESO’s rules concerning the Availability Dispatch Envelope (ADE) for
dispatchable loads. The ADE is the hourly bid quantity made by dispatchable loads into the
IESO’s day-ahead commitment process. The Market Rules and applicable Market Manual
255
Ibid.
PUBLIC VERSION
Products Inc.
256
require that a dispatchable load’s bids into the real-time energy market not exceed 102% of its
ADE. However, the Panel is not aware of any restriction (other than the IESO’s 15 MW
Compliance Deadband) on Thunder Bay consuming energy at a level above its bid quantity.
Thus, in the Panel’s view, the Thunder Bay Facility could have retained a ● MW bid in the realtime market for February 19 through May 11, 2010 and consumed at the level it actually did
while remaining compliant with the ADE rules.
As Resolute notes in its response, consumption at the Thunder Bay Facility during some off-peak
hours was not only higher than ● MW but was also higher than ● MW. The Panel has reexamined the data and agrees that actual consumption was often higher than ● MW (and at times
much lower than ● MW) during the first month of the period in question (or from February 19 to
roughly mid-March 2010). While the Panel has seen nothing that causes it to reconsider its
finding that the increase in bid quantity was for the purpose of increasing CMSC payments, the
Panel reduced its estimate of the amount of CMSC payments obtained through this conduct by
counting payments made in the shorter period from March 16, 2010 to May 11, 2010, inclusive.
b. Finding #9 (ramp down pattern for the Thunder Bay Facility)
Resolute claims that ramping down the load at the Thunder Bay Facility in a single hour was
developed to avoid negative CMSC payments. It cites a September 24, 2009 e-mail as support
for that claim.256 The Panel reviewed that e-mail when drafting this Report, and the Panel still
sees nothing in the e-mail that connects the selected ramp down pattern with negative CMSC
payments. No further records or explanation have been provided by Resolute to substantiate its
stated concern.
Resolute also states that ramping the Thunder Bay Facility down in a single hour “would be of
less risk to DR2 payments as it would take additional load out of the peak period.” The Panel
had already considered this argument but did not accept it as justification for the Facility’s ramp
256
PUBLIC VERSION
Products Inc.
257
down pattern because [Senior Bowater Personnel #5]’s shut down instructions noted that it was
acceptable to drag out the shutdown of auxiliaries into the hour after 7:00 am.257
Resolute’s response to this Finding also states that, just before the Thunder Bay Facility reentered the market as a dispatchable load, two IESO staff members “appeared satisfied” with the
ramp rates Bowater proposed to use for the Facility. The Panel was already aware of this
assertion, having read it in the material supplied in response to the Panel’s Requests for
Information.258 However, Finding #9 does not address Bowater’s ramp rates, but rather relates to
the decision to ramp down the Thunder Bay facility in a single hour rather than delay the ramp
down of auxiliaries to the next hour. Accordingly, the e-mail exchange in question does not
appear to the Panel to be germane to this Finding. There is nothing in Resolute’s response or its
earlier responses to the Panel’s Requests for Information that suggests that IESO staff was
satisfied with the proposed ramping pattern for the Thunder Bay Facility. Even if IESO staff
“appeared satisfied” with Bowater’s decision to ramp down the Thunder Bay Facility in a single
hour, as discussed in section A above the Panel does not consider that to justify or excuse the
market participant behaviour described in this Report.
c. Findings #10 (Thunder Bay Facility ramping down faster than submitted ramp rates)
and #23 (Fort Frances Facility ramping down faster than submitted ramp rates)
Resolute claims that ramping “slightly faster over one or two intervals over the course of a
ramping hour” does not mean that Bowater and Abitibi were guilty of ramping down faster than
submitted ramp rates. Resolute also notes that all of the instances of ramping down identified by
the Panel were within the deadband for dispatch deviations, and further claims that the Panel has
given no consideration to the inherent difficulties of ramping a load such as the Thunder Bay and
Fort Frances Facilities. Resolute’s response implies that a dispatchable load can be found to be
257
258
Responses to RFI, B.2.28, referred to in section 7.4.4 of this Report.
Responses to RFI, B.2.
PUBLIC VERSION
Products Inc.
258
ramping down faster than its submitted ramp rates only if its consumption during an interval falls
below the constrained schedule by an amount greater than the IESO’s Compliance Deadband.
The Panel accepts that precise compliance with dispatch instructions may not be possible in all
situations. However, when a load consistently ramps down faster than its submitted ramp rates,
which the Panel finds was the case here, the Panel has good reason to believe that this ramping
behaviour is not tied exclusively to normal operational problems.
The Panel certainly accepts that the Compliance Deadbands established by the IESO for
generators and loads are appropriate in the context of ensuring the reliable operation of the
electricity system. However, the Panel does not accept that those Compliance Deadbands are
dispositive when assessing whether a market participant is ramping down its facilities faster than
its submitted ramp rates for the purpose of increasing CMSC payments.
Finally, the fact that the deviation between actual consumption and the constrained schedule
during ramp down is small does not make it inherently trivial. As shown in section 7.4.5 of this
Report, small deviations can have a significant impact on CMSC payments when a load is
bidding at $●/MWh.
d. Finding #24 (frequent ramping at the Fort Frances Facility)
Resolute does not dispute that the Fort Frances Facility ramped up and down much more
frequently in the first eight months of 2010 than it did in earlier periods.
Consistent with the information filed in response to the Panel’s Requests for Information, the
response attributes the increased ramping to various operational factors, including the operational
implications of the shutdown of a paper machine in 2009. The response does not provide any
additional information on how and why operational factors required an 85% increase in the
number of self-induced ramps in the first 8 months of 2010 compared to 2009.
As in other parts of its response, Resolute accuses the Panel of attempting to change the Market
Rules retroactively. It states that it “is not aware of a market rule which forbids participants from
PUBLIC VERSION
Products Inc.
259
planning in order to manage and forecast CMSC.” The Panel does not assert in this Report that
such a Market Rule exists. But when “managing” CMSC extends to exploiting market defects
for the benefit of the market participant and at significant cost to the market, then such behaviour
can constitute gaming.
C.
Subsequent Correspondence
By letter dated August 22, 2014, the Panel notified Resolute that it had considered Resolute’s
July 2, 2014 response and altered one aspect of this Report (Finding #8) as a result. The Panel
also advised of the process for finalization of this Report, including the Panel’s intention to
reproduce the entirety of Resolute’s response as an Appendix and to include the Panel’s
comments on certain elements of that response as a separate Appendix. The Panel also identified
the types of information that it agreed to redact from the public version of this Report, and
invited Resolute to provide a redacted version of its July 2, 2014 response if Resolute wished to
request confidential treatment of information contained in that response. Resolute did so and, as
noted above, it is Resolute’s redacted version of its response that appears as Appendix N in the
public version of this Report.
A redacted version of Resolute’s July 2, 2014 response was provided under cover of a letter
dated September 2, 2014. In that letter, Resolute advised that, upon further review, there were
some materials that fall within the scope of the Panel’s October 2010 Requests for Information
that were not discovered until several years after responding to those Requests for Information,
and that it would advise the Panel when its review had been completed. Resolute also provided
certain additional information to the Panel, including an analysis of data that in Resolute’s view
demonstrated that the majority of IESO data respecting dispatch was often inaccurate and could
not be relied upon.
On September 30, 2014, the Panel notified Resolute that, of the supplementary materials
provided on September 2, 2014 that appeared to the Panel to be potentially relevant to its
investigation, the Panel assessment is that they are not inconsistent with the Panel’s findings.
The Panel also advised that the materials did not cause the Panel to change its view regarding the
accuracy of the IESO data used to quantify the impact of the market participants’ conduct.
PUBLIC VERSION
Products Inc.
260
However, the Panel allowed Resolute a further opportunity to identify, with precision, any
specific factual errors in this Report and how the alleged inaccuracies affect the substance of the
Panel’s findings. On October 14, 2014, Resolute advised the Panel that it had completed its
review of the completeness of its responses to the Panel’s Requests for Information to ensure that
the information provided to the Panel was complete and accurate, and provided further
information to the Panel.
D.
Update on the IESO's General Conduct Rule
The Panel notes that amendments to the Market Rules to include the "general conduct rule" were
approved by the IESO Board of Directors on June 12, 2014, and that those amendments have
now come into force.
PUBLIC VERSION

MSP Investigation Report 2015 - CMSC - Abitibi

Transcription

Similar documents

Paycheck Direct Member Purchasing Program

Renova / Nova Scotia Lands Bowater Mersey Mill Project Liverpool

The Prez Sez… Bar Hop January 25 Travel and Adventure Show

Navigating the McKendree Website

We`re ready to serve you from coast to coast H.O. Wolding

Paymentez User How To Guide en

2012MAR6 BGI-PPI Press Release Intl.

We`ll have you up and running in minutes... Parent

GROWERS SUPPLYTS FLEXIBLE FINANCING

CaixaBank